Exposure to Experimental Preeclampsia in Mice Enhances the Vascular Response to Future Injury

Abstract

Cardiovascular disease (CVD) remains the leading killer of women in developed nations. One gender-specific risk factor is preeclampsia (PE), a syndrome of hypertension and proteinuria that complicates 5% of pregnancies. Although PE resolves after delivery, exposed women are at increased long term risk of premature CVD and mortality. Preexisting CVD risk factors are associated with increased risk of developing PE but whether PE merely uncovers risk or contributes directly to future CVD remains a critical unanswered question. A mouse PE model was used to test the hypothesis that PE causes an enhanced vascular response to future vessel injury. A PE-like state was induced in pregnant CD1 mice by overexpressing soluble fms-like tyrosine kinase-1 (sFlt-1), a circulating anti-angiogenic protein that induces hypertension and glomerular disease resembling human PE. Two months post-partum, sFlt-1 levels and blood pressure normalized and cardiac size and function by echocardiography and renal histology were indistinguishable in PE-exposed compared to control mice. Mice were then challenged with unilateral carotid injury. PE-exposed mice had significantly enhanced vascular remodeling with increased vascular smooth muscle cell proliferation (180% increase, P<0.01) and vessel fibrosis (216% increase, P<0.001) compared to control pregnancy. In the contralateral uninjured vessel, there was no difference in remodeling after exposure to PE. These data support a new model in which vessels exposed to PE retain a persistently enhanced vascular response to injury despite resolution of PE after delivery. This new paradigm may contribute to the substantially increased risk of CVD in woman exposed to PE.

Introduction

Despite substantial advances in prevention and treatment strategies, cardiovascular disease (CVD) remains the leading killer of women in the developed world. In the United States, deaths from CVD in women exceed that of men and of the next seven leading causes of death in women combined(1). One gender specific CVD risk factor is a history of preeclampsia (PE), a hypertensive complication during pregnancy(2). PE is one of the most common and serious pregnancy complications effecting five percent of all pregnancies or over 200,000 women annually in the United States. PE is diagnosed by the development of hypertension, proteinuria, and other features of end organ damage during pregnancy such as HELLP syndrome (hemolysis, elevated liver enzymes and low platelets), eclampsia (PE with associated seizures), and even maternal mortality(3). Delivery of the placenta cures the immediate PE episode, yet substantial epidemiologic data demonstrates that affected women have an increased risk of chronic hypertension, premature CVD and death from heart attack or stroke many years postpartum(4–8). Recurrent PE is associated with an even greater CVD risk and women who have had PE before 34 weeks of gestation or PE associated with fetal growth restriction, have a risk of cardiovascular death that is 4–8 times that of women who had a normal pregnancy(2). Overall, the risk of CVD after PE is increased by 2–4 times, an increase comparable to the risk induced by smoking. This strong correlation has prompted the American Heart Association to include a history of PE as an independent risk factor in the 2011 guideline for the prevention of CVD in woman(1).

Substantial progress has been recently made in understanding the pathogenesis of PE. Abnormal placentation results in placental expression of anti-angiogenic factors, including soluble Fms-like tyrosine kinase-1 (sFlt-1), an endogenous inhibitor of vascular endothelial growth factors(3;9). Serum/Plasma levels of sFlt-1 are increased in women with PE, rise weeks before the appearance of clinical manifestations of PE, and correlate with the severity of disease(10–12). High sFlt-1 levels produce endothelial dysfunction which contributes to abnormal vascular tone and hypertension, increased glomerular vascular permeability leading to proteinuria, and consumptive coagulopathy(13). When injected into pregnant rodents, sFlt-1 reproduces systemic endothelial dysfunction resulting in a syndrome that phenocopies human PE, supporting the concept that sFlt-1 is a pathogenic mediator of PE(12;14–16).

In contrast, the etiology for increased CVD risk late after PE is not well understood. Traditional CVD risk factors, including preexisting hypertension, obesity, and diabetes, are also associated with an increased risk for developing PE during pregnancy. As a result of these shared risk factors, it has been proposed that development of PE uncovers a preexisting condition that would have resulted in CVD later in life(17). However, another possibility is that PE-induced vascular damage contributes directly to the future development of maternal CVD. Women with PE demonstrate an increased blood pressure response to angiotensin II that persists after delivery(18) supporting the presence of subclinical vascular dysfunction that could persist after PE. The increased risk of CVD in women with recurrent PE(2) and a decreased incidence of hypertension in the siblings of women with PE who might be expected to be at similar CVD risk to their PE-exposed siblings(19), also support this hypothesis. Recent studies in a mouse model of PE induced by sFlt-1 reveal no detectable differences in blood pressure or in vascular contractile and relaxation function six months after delivery(20) although alterations in circulating proteins persist(21). Thus, whether the predisposition to CVD after PE is caused by exposure to PE or due to shared risk factors remains an important unanswered question(22).

Vascular disease and dysfunction develop in response to damage to the vascular endothelium that may be initiated by exposure to risk factors such as hypertension and diabetes, direct mechanical damage during vascular procedures, and other causes. The damaged endothelium promotes thrombosis and vascular inflammation and also acts in a paracrine fashion on underlying smooth muscle cells (SMC) to activate SMC proliferation and migration that contributes to adverse vascular remodeling(Reviewed in (23)). Vascular remodeling is the process in which vascular hyperplasia and fibrosis compromises vessel function by reducing compliance and inhibiting normal blood flow. Excessive vascular remodeling is an important feature of atherosclerosis and also contributes to the failure of therapies for CVD including stent and vein graft failure and transplant vasculopathy (Reviewed in (24)). Mouse models of vascular remodeling, including the wire carotid injury model(25), have been used for over two decades to explore mechanisms driving vascular remodeling(24). Here we use the classical mouse carotid injury model after exposure to a mouse model of sFlt-1-induced PE to test the hypothesis that prior exposure to PE results in an enhanced vascular response to future vessel injury as this could contribute directly to the enhanced risk of CVD in women exposed to PE.

Methods

Mouse Model of Preeclampsia

All animals were handled in accordance with NIH standards and all procedures were approved by the Institutional Animal Care and Use Committees. Female CD1 mice were randomized to receive either sFlt-1 adenovirus or CMV-null virus at gestational day (GD)9 of pregnancy resulting in hypertension, proteinuria and glomerular endothelial damage resembling human PE as previously described(12;15;16;26;27).

Mouse Carotid Injury Model

Two months post-partum, left common carotid artery endothelial denudation vascular injury was produced with an angioplasty wire and a Bromodeoxyuridine (BrdU) infusion pump was placed subcutaneously. The right carotid artery served as an uninjured control in each mouse. Two weeks after injury, both the carotid arteries were processed for histology and medial area, BrdU positive cells, and fibrosis were each quantified as described(28;29). The overall experimental paradigm is depicted in Figure 1.

Figure 1. Experimental Paradigm.

Pregnant female CD1 mice were randomized to experimental preeclampsia (PE) by injection of sFlt-1 adenovirus (N=8) or control pregnancy by injection of CMV virus (N=6) on gestational day (GD)9. Plasma sFlt1 levels were measured on GD16 and mice were allowed to deliver. Mice were aged for 2 months and plasma sFlt-1 and cardiovascular function by non-invasive tail cuff plethysmography and cardiac ultrasound were measured. The mice then underwent unilateral wire-induced carotid artery injury and implantation of a BrdU infusion pump. Fourteen days later, mice were sacrificed and tissues processed for carotid and renal histology.

For detailed methods descriptions see the Online-only Data Supplement.

Results

A mouse model that reproduces the preeclampsia phenotype during pregnancy with resolution post-partum

To examine the long term vascular response to injury after prior exposure to PE, we used a well validated mouse model of PE(12;20;21) induced by injection of adenovirus expressing sFlt-1 into pregnant mice at GD9 (Figure 1). Pregnant mice were randomized to receive sFlt-1 (experimental PE) or CMV adenovirus (Control) under otherwise identical conditions. sFlt-1 levels were significantly elevated at GD16 (7 days after virus injection) in the experimental PE group compared to Controls (Figure 2A) and correlated with levels found in women with PE(30;31). A second group of mice (N=3 Control, N=5 experimental PE) had telemetry devices inserted prior to pregnancy exclusively to confirm the PE phenotype. In this subset of mice, overexpression of sFlt1 was associated with the characteristics of PE including significantly elevated systolic blood pressure and characteristic renal pathology in late gestation in the PE group similar to what has been described by other groups (Figure S1)(12;15;16;26;27).

Figure 2. Resolution of preeclampsia and normalization of noninvasive measures of cardiovascular function in a mouse experimental PE model.

Mice exposed to sFlt-1 adenovirus (N=8, Prior Preeclampsia or black bars) or CMV virus (N=6, Control or grey bars) during pregnancy were aged for two months after delivery. (A) Plasma sFlt-1 levels were significantly elevated on gestational day (GD)16 of pregnancy (*P < 0.001) and (B) returned to normal levels two months post-partum with no significant difference between groups. (C) Blood pressure was measured two months after pregnancy by tail cuff plethysmography. There was no significant difference in systolic or diastolic pressure between groups. (D) Cardiac function was measured two months after pregnancy by echocardiography. Representative M-mode ultrasound images are shown and the graphs below reveal no significant difference in averaged cardiac chamber dimensions (end diastolic dimension=EDD, end systolic dimension=ESD), cardiac thickness (anterior and posterior wall thickness, AWT and PWT), or fractional shortening between groups. (E) Representative renal histology at the time of animal sacrifice reveals resolution of renal pathology two months post-partum.

The post-PE carotid injury mice (N=6 Control, N=8 Experimental PE) were allowed to deliver pups and then aged for two months. Plasma sFlt-1 levels measured two months post-partum had returned to within the normal range in both groups with no significant difference between the experimental PE and the Controls (Figure 2B). Prior to carotid injury, systolic and diastolic blood pressures were normal and were not significantly different in mice exposed to PE compared to controls (Figure 2C). Since PE has also been associated with cardiac dysfunction in this model and in women(32;33), cardiac structure and function were measured by non-invasive ultrasound two months post-partum. No significant difference was detectable in cardiac chamber dimensions (end diastolic dimension (EDD) and end systolic dimension (ESD)), cardiac thickness (anterior and posterior wall thickness (AWT and PWT)), or fractional shortening (a measure of cardiac contractile function) between mice exposed to PE or Control pregnancy (Figure 2D). Finally, renal histology at the termination of the study reveals resolution of the renal endotheliosis two months after exposure to experimental PE (Figure 2E). Thus, this mouse model phenocopies human PE including the transient hypertension and renal pathology during pregnancy which resolve after delivery along with normalization of plasma sFlt-1 levels and of non-invasive measures of cardiovascular function.

Vascular remodeling in response to injury is enhanced after distant exposure to preeclampsia

To examine whether PE exposure has lasting effects on the vasculature that could contribute to future vascular responses to injury, mice were exposed to unilateral wire-induced endothelial injury two months after either experimental PE or Control pregnancy. Two weeks after carotid injury, the injured and uninjured carotid arteries were isolated and vascular thickness quantified in histological sections stained to visualize elastin fibers. Vessel medial area of the uninjured right carotid arteries is not different in mice exposed to PE compared to Controls (Figure 3, Uninjured). However, when exposed to wire-induced endothelial injury, vessels undergo hypertrophic remodeling and this response is significantly enhanced in the arteries of mice exposed to prior PE (Figure 3, Injured). These data support the concept that although vessel structure appears histologically normal after recovery from PE, when the vessel is exposed to an injury stimulus, the adverse remodeling response is accentuated after PE exposure.

Figure 3. Enhanced remodeling in response to wire injury in mice exposed to prior preeclampsia.

Two months after exposure to experimental preeclampsia (black bars) or Control pregnancy (grey bars), mice were subjected to unilateral wire carotid injury. Vessel medial area was quantified fourteen days after injury in the uninjured and injured carotid arteries. Representative elastin-stained carotid artery sections are shown above at 40X magnification. Bars below represent average medial area for all animals. Scale bar=0.5 mm, N=6 Control, N=8 Prior Preeclampsia. *P<0.001 versus uninjured. NS=not significant.

Enhanced vascular smooth muscle cell proliferation in response to wire injury after prior preeclampsia

One component of the vascular response to endothelial injury is activation of the normally quiescent vascular SMCs to proliferate until the endothelium heals. Bromodeoxyuridine (BrdU) infusion pumps were inserted at the time of injury and proliferation of carotid SMCs during the two weeks after injury was quantified by counting BrdU positive nuclei in the vessel media. As expected, in the absence of an injury stimulus, vascular SMC proliferation is minimal in Control, uninjured vessels. Exposure to prior PE did not alter the very low level of SMC proliferation in uninjured vessels (Figure 4, Uninjured). However, after exposure to wire-induced endothelial injury, SMC proliferation is activated and this proliferative response is significantly enhanced (180% increase, P<0.01) in vessels exposed to prior PE (Figure 4, Injured).

Figure 4. Enhanced smooth muscle cell proliferative response to vascular injury after prior exposure to PE.

Two months after exposure to experimental preeclampsia (black bars) or Control pregnancy (grey bars), mice were subjected to unilateral wire carotid injury along with bomodeoxyuridine (BrdU) infusion to label proliferating cells. Smooth muscle cell (SMC) proliferation was quantified fourteen days after injury in the uninjured and injured carotid arteries. Representative BrdU-stained sections are shown above at 40X magnification. Bars below represent average number of medial BrdU positive SMCs. Scale bar=0.5 mm, N=6 Control, N=8 Prior Preeclampsia. *P<0.001 versus uninjured. NS=not significant.

Increased vascular fibrotic response to injury after distant exposure to preeclampsia

Another component of the vascular injury response is extracellular matrix deposition leading to vascular fibrosis. In the absence of a vascular injury stimulus, medial vessel fibrosis is minimal and not different in vessels exposed to prior PE compared to Control pregnancy (Figure 5, Uninjured). However, the fibrotic response to wire-induced endothelial injury is substantially and significantly increased (216% increase, P<0.001) in vessels exposed to prior PE (Figure 5, Injured).

Figure 5. Enhanced vascular fibrosis in response to wire injury after prior exposure to preeclampsia.

Two months after exposure to experimental preeclampsia (Prior Preeclampsia, black bars) or Control pregnancy (grey bars), mice were subjected to unilateral wire carotid injury. Vascular fibrosis was quantified fourteen days after injury in the uninjured and injured carotid arteries. Representative trichrome-stained sections are shown above at 40X magnification. Bars below represent average medial trichrome pixel area per section. Scale bar=0.5 mm, N=6 Control, N=8 Prior Preeclampsia. *P<0.001 versus uninjured. NS=not significant.

Discussion

A mouse model that faithfully reproduces the syndrome of PE was used to explore whether there is a direct causative effect of PE on future vascular remodeling in response to injury. Mice without underlying cardiovascular risk factors were randomly assigned to exposure to PE or normal pregnancy. Since PE is associated with hypertension and with increased risk of cardiac systolic dysfunction both in women and in this mouse model of PE(32;33), resolution of hypertension and normal cardiac function after delivery was confirmed by non-invasive methods similar to those used to follow women after PE. Two months post-partum, mice were challenged with a vascular injury stimulus and mice exposed to prior PE had a substantially increased vascular remodeling response. The increase in vessel remodeling was due to both an increase in SMC proliferation and enhanced vessel fibrosis in injured vessels exposed to prior PE. These data support a new paradigm in which vessels exposed to PE during pregnancy retain an enhanced vascular response to future injury (Figure 6). The lack of changes in vascular structure in uninjured vessels after exposure to PE supports the idea that an additional vascular injury stimulus is necessary and synergistic with previous PE exposure. Pre-pregnancy cardiovascular risk factors are associated with increased risk of PE although the mechanism for this predisposition is not clear. In this new paradigm, continued exposure to hypertension, obesity or diabetes, or the development of these or other cardiovascular risk factors (smoking, dyslipidemia) later in life which damage the vasculature, would then elicit an enhanced remodeling response in women with a history of PE. Damage to the endothelium, by denudation and/or CVD risk factors, results in endothelial dysfunction with decreased nitric oxide production that activates the normally quiescent SMCs to proliferate and migrate into the intima(23) (as seen in Figure 4 in the injured vessel exposed to prior PE). This pathology contributes to atherosclerotic plaque development, stent restenosis after vascular interventions, vein graft failure, and transplant vasculopathy(24). The associated vascular hypertrophy and fibrosis also contributes to increased vascular stiffness which has been demonstrated in woman with a history of PE(34) and is associated with increased risk of CVD(35). Therefore, this new mechanism may be contributing to the substantially increased risk of future CVD in women exposed to prior PE.

Figure 6. New paradigm for enhanced cardiovascular risk in women exposed to preeclampsia.

This study demonstrates in a mouse model that preeclampsia results in persistent enhanced vascular responsiveness to future injury. These data support a new model in which the vasculature of women exposed to PE is hyper-responsive to injury stimuli long after resolution of PE. Vascular injury may result from persistence of pre-pregnancy cardiovascular disease (CVD) risk factors that are known to predispose to PE as well as from new CVD risk factors, mechanical injury, and other sources of vascular damage as these women age. Enhanced vascular remodeling may contribute to the increased risk of future hypertension and atherosclerosis in this population and could explain the resulting increased risk of heart attack, stroke and cardiovascular death in woman with a history of PE.

Although non-invasive and histologic measures of cardiac and vascular structure and function are indistinguishable after PE compared to normal pregnancy in this animal model, these data suggest that vascular physiological changes persist that exacerbate the vascular response to injury. Whether PE exposure enhances other vascular responses to endothelial injury such as vascular inflammation and hypercoaguability remain to be determined using additional models of CVD. Moreover, the detailed molecular mechanism by which PE exposure leads to such sustained changes in the vasculature remains to be elucidated. Vascular remodeling involves multiple cell types with the intact healthy endothelium acting in a protective fashion to prevent excessive SMC proliferation. Thus, exposure to PE could enhance vascular remodeling in this model by slowing the rate of endothelial healing and/or by directly affecting SMC proliferative capacity. There are several pathways that could be implicated in the mechanism that warrant additional study. Vascular contractile responses to renin-angiotensin-aldosterone system activation are enhanced during and just after a pregnancy complicated by PE both in animal models and in humans(18;36;37). The mechanism for enhanced angiotensin II sensitivity and whether it persists long after delivery remain to be examined. Although one study in the same mouse PE model revealed no difference in vascular contractile responsiveness to adrenergic agonists, thromboxane, or serotonin six months post-partum, responsiveness to renin-angiotensin-aldosterone system agonists was not explored(20). In the same model, persistent alterations in the plasma proteome were identified six months after experimental PE despite normalization of serum sFlt-1 levels(21). Such persistent alterations in plasma proteins could directly contribute to future vascular injury responses and also may be studied as novel biomarkers of cardiovascular risk in this population. Increased circulating autoantibodies to the angiotensin receptor that occurs in PE could contribute to long term cardiovascular risk(38). Other possibilities include epigenetic changes in the vasculature that occur as a consequence of vascular injury that may lead to long term cardiovascular disease. These results support the need for future basic research exploring the underlying mechanism for PE-induced cardiovascular damage.

Based on the substantial clinical data supporting prior PE as an important gender-specific cardiovascular risk factor for heart attack, stroke, and cardiovascular death in women(2), American Heart Association guidelines recommend screening woman based on pregnancy history. However, at this time, no specific therapies are available to prevent these outcomes in woman identified to have increased risk by such screening. Future clinical studies are needed to address whether inclusion of a history of PE into risk prediction algorithms might be beneficial in determining treatment goals for traditional risk factor modifications including cholesterol lowering and blood pressure targets in women. Based on this study, it might be postulated that woman exposed to prior PE would be at higher risk of in-stent-restenosis after percutaneous vascular procedures or of vein graft failure after bypass surgery. This remains to be explored in epidemiologic studies but if confirmed, it might be beneficial to choose coated stents or arterial conduits in this patient population. In addition, the current study supports the need for further exploration of the molecular mechanisms by which PE directly contributes to the pathophysiology of CVD. Elucidation of the mechanisms could identify novel gender-specific treatment targets to prevent the progression of CVD in women identified early as having increased risk based on their pregnancy history. Such novel therapies could substantially reduce CVD morbidity and mortality in this population of woman at high risk.

Perspectives

Exposure to preeclampsia during pregnancy is associated with increased future risk of CVD in women. This has been attributed to pre-existing CVD risk factors in patients with PE. This study demonstrates that in a mouse model without pre-existing risk factors, PE exposure during pregnancy potentiates the adverse vascular remodeling response to injury later in life. The vascular injury response in mice is enhanced after PE despite complete normalization of non-invasive cardiovascular parameters after delivery, as in women with PE. Moreover, examination of uninjured vessels reveals that in the absence of a vascular injury stimulus, there is no difference in vascular remodeling after PE. This supports a new paradigm in which PE causes changes in vascular physiology that enhance the response to future vascular damage that may be mediated by pre-existing and/or new risk factors to which women are exposed after PE. Since woman can be identified very early as being at high risk due to their exposure during pregnancy, this new understanding sets the stage for basic and clinical studies to determine treatments that can prevent the rapid progression of CVD in these high risk women either by aggressively treating cardiovascular risk factors or by modulating the underlying physiology.

Novelty and Significance.

1. What Is New?

• This study demonstrates for the first time that exposure to PE during pregnancy enhances the adverse vascular remodeling response to future injury in a mouse model.

• PE exposure enhances the vascular injury response independent of pre-existing cardiac risk factors in this model.

2. What Is Relevant?

• Women exposed to PE during pregnancy have an increased risk of heart attack, stroke, and cardiovascular death that is comparable to a history of smoking.

• While this has been attributed to shared risk factors for PE and CVD, this study suggests that PE may also cause persistent vascular dysfunction that directly contributes to future risk of CVD and may lead to future novel gender-specific therapeutic opportunities.

3. Summary

• This study used a mouse model of PE to explore potential mechanisms for the substantial increased risk of CVD in women with a history of PE. PE induced in mice by sFlt-1 injection reproduces the human syndrome during pregnancy with normalization of cardiovascular structure and non-invasive measures of function two months post-partum. Mice exposed to PE demonstrate enhanced vascular responsiveness to future injury. This new paradigm may contribute to the substantially increased risk of CVD in woman exposed to PE.