Both endothelial mineralocorticoid receptor expression and hyperleptinemia are required for clinical characteristics of placental ischemia in mice

Abstract

One of the initiating events in preeclampsia (PE) is placental ischemia. Rodent models of placental ischemia do not present with vascular endothelial dysfunction, a hallmark of PE. We previously demonstrated a role for leptin in endothelial dysfunction in pregnancy in the absence of placental ischemia. We hypothesized that placental ischemia requires hyperleptinemia and endothelial mineralocorticoid receptor (ECMR) expression to induce PE-associated endothelial dysfunction in pregnant mice. We induced placental ischemia via the reduced uterine perfusion pressure (RUPP) procedure in pregnant ECMR-intact (ECMR^+/+) and ECMR deletion (ECMR^−/−) mice at gestational day (GD) 13. ECMR^+/+ RUPP pregnant mice also received concurrent leptin infusion via miniosmotic pump (0.9 mg/kg/day). RUPP increased blood pressure via radiotelemetry and decreased fetal growth in ECMR^+/+ pregnant mice. Both increases in blood pressure and reduced fetal growth were abolished in RUPP ECMR^−/− mice. Placental ischemia did not decrease endothelial-dependent relaxation to acetylcholine (ACh) but increased phenylephrine (Phe) contraction in mesenteric arteries of pregnant mice, which was ablated by ECMR deletion. Addition of leptin to RUPP mice significantly reduced ACh relaxation in ECMR^+/+ pregnant mice, accompanied by an increase in soluble FMS-like tyrosine kinase-1 (sFlt-1)/placental growth factor (PLGF) ratio. In conclusion, our data indicate that high leptin levels drive endothelial dysfunction in PE and that ECMR is required for clinical characteristics of hypertension and fetal growth restriction in placental ischemia PE. Collectively, we show that both ECMR and leptin play a role to mediate PE.

NEW & NOTEWORTHY Leptin is a key feature of preeclampsia that initiates vascular endothelial dysfunction in preeclampsia characterized by placental ischemia. Endothelial mineralocorticoid receptor (ECMR) deletion in placental ischemia protects pregnant mice from elevations in blood pressure and fetal growth restriction in pregnancy. Increases in leptin production mediate the key pathological feature of endothelial dysfunction in preeclampsia in rodents. ECMR activation contributes to the increase in blood pressure and fetal growth restriction in preeclampsia.

INTRODUCTION

Preeclampsia (PE) is a leading cause of maternal mortality in the United States affecting ∼5–7% of the total pregnancies (1–3). PE is a pregnancy-induced hypertension disorder that is diagnosed after the 20th week of gestation by new onset hypertension, alongside a symptom of placental or renal dysfunction (4). PE confers a significant maternal and fetal clinical burden with long-term cardiovascular complications for years following delivery (2). Although the causes and risk factors for the development of PE such as genetic predisposition, previous PE pregnancy, advanced maternal age, and nulliparity (5–7) indicate diverse causes of disease, the pathogenesis in mid-late gestation is fairly uniformly characterized by maternal endothelial dysfunction (8–10).

In many PE cases, an initiating event of disease is placental ischemia (11), which has led to the development of animal models recapitulating placental ischemia in rodents. Experimentally, surgical reduction of uteroplacental perfusion to the placenta, termed reduced uterine perfusion pressure (RUPP), results in many of the characteristics of PE in rats (12, 13). In mice, the RUPP model is newer and less well described than in rats but has been demonstrated to increase blood pressure and promote fetal growth restriction (14–16). Endothelial dysfunction is a key characteristic of PE found both before the onset of symptoms and at the time of presentation after the 20th week of gestation (17–20). However, in contrast to the high association of impaired endothelial relaxation responses found in PE women (10), rodent PE models, including the RUPP, do not consistently demonstrate disruption of endothelial relaxation in the systemic vasculature (21–24). Therefore, there is likely a PE factor in humans that is absent from these rodent models mediating vascular endothelial dysfunction.

In nonpregnant female mice, endothelial mineralocorticoid receptor (ECMR) activation is a crucial sex-specific mechanism whereby obesity and salt sensitivity induce endothelial dysfunction (25–27) and we have demonstrated that ECMR increases in pregnancy (28). In addition, we recently showed that ECMR deletion protects pregnant mice from fetal growth restriction, hypertension, and endothelial dysfunction in a mouse model of hyperleptinemia of pregnancy (29). Therefore, we sought to test in this study whether leptin is a crucial mechanism whereby placental ischemia induces systemic endothelial dysfunction and whether ECMR plays a role in placental ischemia-induced PE.

MATERIALS AND METHODS

Experimental Animals

All protocols were approved by the Institutional Animal Care and Use Committee (Protocol No. 2011-0108) of Augusta University. Mice were kept at room temperature and maintained on a 12-h:12-h light/dark cycles with free access to standard rodent chow. The generation of female mice with an endothelial-specific mineralocorticoid receptor (MR) knockout (ECMR^−/−) has been previously described (30). Endothelial cell-specific MR deletion (flox/flox) was achieved through cross with vascular endothelial-cadherin Cre recombinase. Timed pregnancies in female mice of either ECMR^−/− or their wild-type Cre negative littermates (ECMR^+/+) were performed with the assessment of gestational day 1 (GD1) through vaginal plug visualization and weights were tracked.

Placental Ischemia Induction by Reduced Uterine Perfusion Pressure Model

On GD13, ECMR^−/− or ECMR^+/+ mice were randomized to either placental ischemia via the reduced uterine perfusion pressure (RUPP) model or sham as described (14). Briefly, for the RUPP surgery, an abdominal incision was made, and the abdominal aorta was exposed and cleaned of any periadventitial fat. With the aid of a metal rod parallel to the aorta, a ligature was tied around the aorta before the metal rod being swiftly removed resulting in reduced blood flow. Two silver clips were clamped on both the right and left ovarian vascular arcade leaving a 0.6-mm gap for reduced blood flow (14). For the sham surgeries, abdominal incision was made, and blood vessels were manipulated in a similar manner as in the RUPP surgery without any ligatures or clamps. The abdominal incisions were sutured, mice were provided with buprenorphine analgesic and monitored for recovery before being housed individually. The mice were euthanized on GD18.

Conscious Blood Pressure Measurements

For blood pressure (BP) measurement, a specific cohort of mice was implanted with radiotelemeters via catheter implantation in the right carotid arteries (PA-C10, DSI Sciences International, St. Paul, MN) for continuous measurement. Animals were allowed 7 days of recovery from the surgery before baseline blood pressure and heart rate values were obtained for 3 days before breeding. GD15–GD18 blood pressures and heart rates were then recorded and analyzed as described by Faulkner and colleagues (29, 31). n values vary between groups because of instances of adverse outcomes of sham or RUPP surgery (notably hindlimb ischemia) requiring euthanasia, failure to breed successfully, or telemeter failure in the pregnancy recordings.

Vascular Reactivity

Thoracic aorta and second-order mesenteric resistance arteries from a specific cohort of pregnant mice at GD18 were identified, cleaned of perivascular vascular adipose, and mounted on wire myography (DMT, Ann Harbor, MI) as described (27, 32) for assessment of vascular function. For each animal (each n of the study) four rings were mounted on a myograph, the mean of the measures of these four rings is the n depicted in the figure. Vessels were normalized using the DMT normalization module and left to equilibrate for 30 min before KCl for maximum depolarization contractions. Aorta was preconstricted to serotonin whereas mesenteric arteries were preconstricted to phenylephrine (Phe) before being assessed for endothelial-dependent relaxation by responses to acetylcholine (ACh) and endothelial-independent relaxation responses to sodium nitroprusside (SNP) (log. curves 1 nM–30 µM concentrations). Contribution of nitric oxide synthase (NOS) was determined by preincubation with NOS inhibitor N^ω-nitro-l-arginine methyl ester (l-NAME, 100 μM, 20-min preincubation) followed by concentration-response curves to ACh. In aorta only, we also measured the contribution of different endothelial-derived relaxation factors cyclooxygenase-derived prostanoids (COX, indomethacin), and endothelial-derived hyperpolarizing factors [EDHF, TRAM34 plus apamin (25)] via preincubation (1 µM each, 20-min preincubation) with respective inhibitors followed by concentration-response curves to ACh. Constriction response curves to Phe were also performed and expressed as a percentage of KCl. Responses to ACh and SNP were expressed as a percentage of preconstriction (32). Maximal relaxation and half-maximal effective concentration (EC₅₀) were also calculated for mesenteric artery concentration-response curves. Data were acquired, recorded, and analyzed using LabChart software (AD Instruments, Colorado Springs, CO). n values vary between groups because of instances of adverse outcomes of sham or RUPP surgery (notably hindlimb ischemia) requiring euthanasia or failure to breed successfully.

Fetal Assessment

On GD18, all mice from both blood pressure and vascular reactivity cohorts were weighed before being euthanized under isoflurane anesthesia, and blood was collected. An abdominal midline incision was made to expose both right and left uterine horns. From each uterine horn, the litter size and viable pup counts were counted and recorded. Fetal demise was recorded as a percentage of the total litter size considering reabsorbed pups that were observed at the time of RUPP or sham surgery. Viable fetuses and their corresponding placentas were extracted and weighed, and placental efficiency was calculated. Tissues were quickly snap frozen in liquid nitrogen and stored at −80°C, and plasma was isolated by centrifugation.

Leptin Infusion

On GD11–GD18, a separate cohort of pregnant mice was implanted with subcutaneous osmotic minipumps (ALZET, Cupertino, Calif; model 1007 D, 0.5 μL/h) containing leptin (CYT-351, Prospec, Rehovot, Israel) released at a dose of 0.9 mg/kg/day as previously described (29).

Western Blot Analysis

Whole placenta was homogenized in tissue protein extraction reagent complemented with protease and phosphatase inhibitors before protein quantification using bicinchoninic acid (BCA) kit as described (31). Protein was run on SDS-PAGE gels before being transferred to a PVDF membrane for probing with antibodies and subsequent visualization by enhanced chemiluminescence (Cytiva Cat. No. RPN2236). The following antibodies were used phospho-endothelial nitric oxide synthase (p-eNOS) ser1177, (BD Transduction, Cat. No. 612393), at a concentration of 1:500, eNOS (Cat. No. 610297) concentration of 1:500; and heat shock protein 90 (HSP90) (Cat. No. 610419, BD Transduction Laboratories) concentration of 1:1,000. Whole blot images with marker are located in Supplemental Fig. S1. Data were normalized, and band intensity was quantified by ImageJ software.

Measurements of Plasma sFlt-1, PLGF-2, and Leptin

Plasma sFlt-1 (Cat. No. MVR100, R&D Systems, Minneapolis, MN), PLGF-2 (Cat. No. MP200, R&D Systems, Minneapolis, MN), and leptin (Cat. No. EZML-82K, Millipore Sigma, Danvers, MA) were quantified by commercially available enzyme-linked immunosorbent assays (ELISA) as per the manufacturer’s instructions.

Statistical Analysis

All data sets were expressed as means ± SE and analyzed using Graphpad Prism 9 (GraphPad Software, La Jolla, CA). Multiple data sets used two-way ANOVA, with an assessment of intergroup and interaction effect, or one-way ANOVA. Vascular dose-response curves were analyzed with two-way ANOVA with repeated variable measures. Sidak or Fisher’s post hoc test followed all ANOVA for multiple comparisons. For all comparisons, P values < 0.05 were considered significant.

RESULTS

Placental Ischemia Induced by RUPP Increased BP Significantly While ECMR Deletion Attenuated the Increases in BP

Conscious blood pressure (BP) recordings as well as heart rate were recorded before and during pregnancy GD15–18 in ECMR^+/+ and ECMR^−/− sham or RUPP. BP and heart rate measures in ECMR^+/+ and ECMR^−/− sham and RUPP pregnant mice did not differ before pregnancy (Supplemental Fig. S2). RUPP significantly increased mean arterial pressure (MAP, Fig. 1A) and diastolic blood pressure (DBP, Fig. 1B) in pregnant mice with an intact ECMR compared with their sham. Systolic blood pressure (SBP) did not significantly increase in ECMR^+/+ + RUPP compared with ECMR^+/+ + sham, although this did achieve a P value of 0.06 (Fig. 1C). No effect of RUPP was found for heart rate (HR) in ECMR^+/+ pregnant mice (Fig. 1D). ECMR deletion attenuated the ability of placental ischemia by RUPP to increase MAP and DBP (Fig. 1, A and B). RUPP had no effect on increasing or decreasing SBP and HR in ECMR^−/− pregnant mice, however, it is important to note that the P value of the multiple comparison analysis comparison of ECMR^+/+ + RUPP versus ECMR^−/− + RUPP for SBP achieved a P value of 0.05. Therefore, in the absence of ECMR expression, placental ischemia in pregnant mice does not increase BP.

Figure 1.

Placental ischemia increases blood pressure in endothelial mineralocorticoid receptor (ECMR) intact, but not knockout, pregnant mice. Sham or reduced uterine perfusion pressure (RUPP) procedure was performed at gestational day (GD) 13 on both ECMR^+/+ and ECMR^−/− pregnant mice that had been implanted with radio telemeters for conscious blood pressure measurement before mating. Measurement was summarized in pregnancy across GD15–18 for mean arterial pressure (MAP; A), diastolic blood pressure (DBP; B), systolic blood pressure (SBP; C), and heart rate (D). Means ± SE, two-way ANOVA with Fisher’s post hoc test for multiple comparisons. *P < 0.05 and **P < 0.01.

Placental Ischemia Does Not Induce Endothelial Dysfunction in Second Order Mesenteric Resistance Arteries or Thoracic Aorta

Vascular function was assessed by wire myography in second-order mesenteric resistance arteries. Placental ischemia via RUPP did not significantly decrease endothelial-dependent relaxation responses to ACh in resistance arteries (Fig. 2A) of pregnant mice at GD18. Preincubation of the mesenteric arteries with NOS inhibitor l-NAME decreased endothelial-dependent relaxation across all groups equally with no significant difference observed with RUPP or ECMR deficiency (Fig. 2B). Neither RUPP nor ECMR deletion resulted in decreases in endothelial-independent relaxation responses to SNP (Fig. 2C). RUPP significantly increased smooth muscle contraction to Phe in ECMR-intact mice, but ECMR deficiency blunted RUPP-induced Phe response increases in resistance arteries (Fig. 2D). No statistically significant differences to KCl contraction were observed across all groups (Fig. 2E). We further calculated EC₅₀ and maximal relaxation/constriction to ACh, SNP, and Phe as depicted in Table 1. Similar to our two-way ANOVA analysis of concentration-response curves, we found no significant increase in EC₅₀ of ACh- or SNP-mediated relaxation in ECMR^+/+ + RUPP compared with ECMR^+/+ + sham, nor decrease in maximal relaxation between these two groups. ECMR deletion had no effect to decrease EC₅₀ or increase maximal relaxation of ACh or SNP in either sham or RUPP pregnant mice. We found that EC₅₀ and maximal constriction of Phe-mediated constriction significantly increased in ECMR^+/+ + RUPP compared with ECMR^+/+ + sham, similar to what was observed with the two-way ANOVA of the concentration-response curve, thereby indicating an increased vascular sensitivity to Phe-mediated constriction conferred by RUPP in the presence of intact ECMR. However, ECMR deletion ablated the ability of RUPP to increase Phe sensitivity in the mesenteric vasculature of pregnant mice.

Figure 2.

Placental ischemia does not induce endothelial dysfunction. Upon euthanasia gestational day (GD) 18, mesenteric artery beds and placenta tissue were collected from endothelial mineralocorticoid receptor (ECMR)^+/+ and ECMR^−/− pregnant mice previously subjected to sham or reduced uterine perfusion pressure (RUPP) procedure on GD13. We measured vascular function by wire myography in 2nd-order mesenteric resistance arteries. Here we report vascular relaxation (following preconstriction) to acetylcholine (ACh; A) and ACh with the presence of N^ω-nitro-l-arginine methyl ester (l-NAME) preincubation [expressed both by the curve and area under the curve (AUC); B]. We also report relaxation to sodium nitroprusside (SNP; C), as well as constriction to phenylephrine (Phe; D) and KCl (E). Means ± SE, two-way ANOVA with repeated measures for vascular reactivity. For vascular data, n = 8 ECMR^+/+ + sham, n = 6 ECMR^+/+ + RUPP, n = 7 ECMR^−/− + sham, and n = 5 ECMR^−/− + RUPP. *P < 0.05.

Table 1.

Analysis of EC₅₀ and maximal relaxation/constriction responses

	ECMR +/+			ECMR −/−
	+Sham	+RUPP	+RUPP + leptin	+Sham	+RUPP
Maximum
ACh	97.75 ± 2.60	90.83 ± 6.60	66.60 ± 9.27†	84.30 ± 5.14	89.20 ± 8.22
Phe	136.70 ± 11.43	278.50 ± 80.74*	162.9 ± 40.77	225.00 ± 64.95	153.40 ± 14.60
SNP	116.60 ± 8.56	107.20 ± 4.57	95.50 ± 2.63	96.86 ± 3.20	100.80 ± 0.58
EC50
ACh	−7.11 ± 0.19	−8.42 ± 0.33	−6.97 ± 0.19	−7.28 ± 0.31	−6.36 ± 0.86
Phe	−5.37 ± 0.10	−6.03 ± 0.14*	−5.75 ± 0.19	−5.80 ± 0.12	−6.02 ± 0.16
SNP	−6.94 ± 0.17	−7.21 ± 0.12	−7.95 ± 0.08	−7.09 ± 0.29	−7.39 ± 0.29

Values are means ± SE. Maximal relaxation (ACh, maximum SNP) is a percentage of preconstruction. Maximal constriction (maximum Phe) is a percentage of KCl constriction. EC₅₀ is the log-transformed half-maximal effective concentration (EC₅₀). One-way ANOVA was used. *P < 0.05 vs. same-strain sham. †P < 0.05 vs. ECMR^+/+ + RUPP.

In addition, neither RUPP nor ECMR deletion decreased endothelial-dependent relaxation responses to ACh in the thoracic aorta (Supplemental Fig. S3A). l-NAME preincubation decreased overall relaxation to ACh-mediated relaxation in the aorta with no differences with RUPP or ECMR deletion (Supplemental Fig. S3B). To investigate the mechanisms contributing to endothelial dysfunction, the aorta was preincubated with l-NAME and indomethacin, a cyclooxygenase (COX)1 and COX2 inhibitor (indomethacin). We observed that indomethacin did not decrease endothelial relaxation to ACh more so in response to RUPP or ECMR deletion (Supplemental Fig. S3C). Our data also revealed no difference in endothelial-dependent relaxation to ACh with RUPP or ECMR deletion with aorta preincubated with l-NAME and TRAM34 plus apamin (Supplemental Fig. S3D), an EDHF inhibitor cocktail (25). There were no decreases in aorta endothelial-independent relaxation responses to SNP (Supplemental Fig. S3E), as well as no increases in contraction to Phe (Supplemental Fig. S3F) and KCl contraction (Supplemental Fig. S3G) with RUPP or ECMR deletion compared with control. Therefore, in both the resistance vasculature and the conduit vasculature, placental ischemia in mice does not induce reductions in endothelial-dependent dilation. However, in the resistance vasculature only, vascular contractile responses to phenylephrine increase in RUPP mice compared with sham, which is ablated by ECMR deletion.

Placental Ischemia Does Not Decrease, and ECMR Deletion Does Not Increase, eNOS Phosphorylation in Placenta of Pregnant Mice

We used placenta lysates to quantify the level of eNOS phosphorylation (s1177) as well as total eNOS protein expression. ECMR deficiency and RUPP did not reduce the ratio of phosphorylated eNOS compared with total eNOS levels (Fig. 3A). These data, alongside that of Fig. 2, which depicts no change in vascular endothelial relaxation responses in RUPP mice compared with sham, indicate that placental ischemia alone in pregnant mice does not reduce placental eNOS phosphorylation.

Figure 3.

Placental ischemia does not decrease phosphorylation of endothelial nitric oxide synthase (eNOS) or increase leptin levels in pregnant mice. Placental protein expression of phospho-eNOS (p-eNOS) and total eNOS did not decrease following placental ischemia or endothelial mineralocorticoid receptor (ECMR) deletion (A). Plasma levels of leptin were measured following sham or placental ischemia in ECMR^+/+ and ECMR^−/− pregnant mice via enzyme-linked immunosorbent assay (ELISA) from the collection taken at gestational day (GD) 18 (B). Means ± SE, two-way ANOVA with Fisher’s post hoc test for multiple comparisons for protein expression.

Placental Ischemia Does Not Increase Plasma Leptin Levels

We have found, similar to others (33), in contrast to humans placental leptin levels are undetectable in placentas of pregnant mice (mRNA expression of mouse ob/ob leptin gene, not reported because of lack of data). In accordance, we measured leptin levels following sham or RUPP in ECMR^+/+ and ECMR^−/− pregnant mice in plasma collected at GD18. Our results showed that neither RUPP nor ECMR deletion induced an increase in plasma leptin levels (Fig. 3B). Therefore, potentially because of no placental leptin production in the species, placental ischemia does not induce increases in leptin in pregnant mice.

Leptin Infusion Induces Endothelial Dysfunction following Placental Ischemia in ECMR-Intact Mice

We measured vascular function in mice given leptin in addition to RUPP surgery to induce placental ischemia. Our results show that leptin infusion decreased endothelial-dependent relaxations to ACh in mesenteric arteries in pregnant ECMR^+/+ + RUPP mice compared both to ECMR^+/+ + sham and ECMR^+/+ + RUPP (Fig. 4A), indicating endothelial dysfunction. Leptin reduced ACh-mediated relaxation in mesenteric vessels preincubated with l-NAME more so than both sham and RUPP mice without leptin infusion, indicating that leptin significantly reduced NOS-mediated endothelial relaxation in pregnant RUPP mice (Fig. 4B). Leptin infusion in RUPP pregnant mice did not significantly reduce SNP-mediated vascular relaxation (Fig. 4C). However, leptin reduced Phe-mediated (Fig. 4D), but not KCl-mediated (Fig. 4E) contractile responses in mesenteric arteries of RUPP pregnant mice. Similar to our results with two-way ANOVA of concentration-response curves, in Table 1 we found that leptin addition to RUPP in ECMR^+/+ pregnant mice significantly reduced maximal relaxation to ACh, without impacting relaxation to SNP, when compared with both ECMR^+/+ + sham and ECMR^+/+ + RUPP. Interestingly, leptin also increased EC₅₀ of SNP-mediated relaxation in ECMR^+/+ RUPP + leptin compared with ECMR^+/+ + RUPP, indicating a heightened sensitivity to SNP in addition to protecting from heightened Phe-mediated constriction. Therefore, leptin induces vascular endothelial dysfunction in mice with placental ischemia and also induces reduced sensitivity of the mesenteric vasculature to α₁-receptor agonism by Phe.

Figure 4.

Leptin infusion induces endothelial dysfunction in endothelial mineralocorticoid receptor (ECMR) intact pregnant mice with placental ischemia. Vascular relaxation responses in 2nd-order mesenteric arteries of ECMR^+/+ + sham, ECMR^+/+ + reduced uterine perfusion pressure (RUPP), and ECMR^+/+ + RUPP + leptin pregnant mice [gestational day (GD) 18]. Here we report vascular relaxation (following preconstriction) to acetylcholine (ACh; A) and ACh (B) with the presence of N^ω-nitro-l-arginine methyl ester (l-NAME) preincubation [expressed both by the curve and area under the curve (AUC)]. We also report relaxation to sodium nitroprusside (SNP; C) and constriction to phenylephrine (Phe; D) and KCl (E). n = 8 ECMR^+/+ + sham, n = 6 ECMR^+/+ + RUPP, and n = 5 ECMR^+/+ + RUPP + leptin. Means ± SE, two-way ANOVA with repeated measures for vascular reactivity, and one-way ANOVA with Fisher’s post hoc test for multiple comparisons for bar graphs. *P < 0.05.

Placental Ischemia-Induced Fetal Growth Restriction and Reduced Placental Efficiency in Pregnant Mice, Which Was Prevented by ECMR Deletion

RUPP significantly reduced GD18 pup weights in ECMR^+/+ pregnant mice at GD18 whereas no decrease in pup weights was observed following RUPP in ECMR^−/− mice (Fig. 5A). Neither RUPP nor ECMR deletion altered placenta weights in pregnant mice (Fig. 5B). However, placenta efficiency, calculated as a ratio of placenta weight to pup weight, was significantly decreased by RUPP in intact ECMR mice with no effect of RUPP to decrease placental efficiency in ECMR^−/− pregnant mice (Fig. 5C). Furthermore, neither RUPP nor ECMR deletion result in significant decreases in fetal reabsorptions (Fig. 5D), litter sizes (Fig. 5E), or maternal body weight both including fetal/placental weight (Fig. 5F) and excluding it (Fig. 5G). ECMR^+/+ + RUPP mice with leptin infusion did not demonstrate reductions in pup weight compared with ECMR^+/+ + sham or ECMR^+/+ + RUPP mice, however, did demonstrate significantly reduced placental weight and an increased placental efficiency (Fig. 5, A–C). Leptin infusion did not significantly alter fetal demise or litter size in ECMR^+/+ + RUPP pregnant mice (Fig. 5, D and E) and importantly, leptin had no effect to reduce maternal body weight with or without the weight of fetus/placenta accounted (Fig. 5, F and G). Therefore, ECMR deletion rescued fetal growth in RUPP mice. Leptin did not restore pup weights in ECMR^+/+ + RUPP mice but did reduce placental size and increase efficiency indicating a discrepant effect of leptin on placental function compared with placental ischemia alone.

Figure 5.

Placental ischemia reduces pup weights and reduces placental efficiency in endothelial mineralocorticoid receptor (ECMR) intact, but not ECMR knockout, pregnant mice. On gestational day (GD) 13, ECMR^+/+ and ECMR^−/− pregnant mice were subjected to sham or reduced uterine perfusion pressure (RUPP) procedure, and a group of ECMR^+/+ + RUPP mice was also treated with leptin infusion. On GD18 fetal outcomes were summarized by fetal weights (A), placental weight (B), placental efficiency calculated as pup-to-placenta ratio (C), fetal demise calculated as resorptions %liter (D), total litter size (E), maternal body weight at GD18 (F), and maternal body weight with sum of placenta-to-pup weights subtracted (G). Means ± SE, one-way ANOVA with Fisher’s post hoc test for multiple comparisons. Number of litters assessed were as follows: ECMR^+/+ + sham, n = 11; ECMR^+/+ + RUPP, n = 10; ECMR^−/− + sham, n = 8; ECMR^−/− + RUPP, n = 9; and ECMR^+/+ + RUPP + leptin, n = 5. *P < 0.05, **P < 0.01, and ***P < 0.001.

Plasma sFlt-1-to-PLGF Ratio Is Increased by Leptin in RUPP Mice

In Fig. 6, we demonstrate that neither RUPP nor ECMR deletion had a significant effect to increase sFtl-1 levels or decrease PLGF levels in the plasma of pregnant mice, which was also not significantly altered by leptin (Fig. 6, A and B). Other groups have similarly reported that RUPP in mice does not consistently increase plasma sFlt-1 levels (15). We found that while the individual measures of sFlt-1 and PLGF were not altered in plasma across groups, sFlt-1/PLGF ratio did significantly increase in ECMR^+/+ + RUPP mice infused with leptin compared with ECMR^+/+ + sham and ECMR^+/+ + RUPP. Therefore, at least in mice, leptin infusion increases the severity of the placental ischemia by inducing this clinical marker of severe PE symptoms in elevated sFlt-1/PLGF.

Figure 6.

Leptin infusion in the presence of placental ischemia increased plasma soluble functional magnetic stimulation (FMS)-like tyrosine kinase-1 (sFlt-1)-to-placental growth factor-2 (PLGF-2) ratio. At gestational day (GD) 18, we measured the levels of sFlt-1 and PLGF (PLGF-2 assay) in the plasma of our pregnant mice in all groups. Here we report levels of sFlt-1 (A), PLGF (B), and the calculated sFlt-1-to-PLGF ratio (C). Means ± SE, one-way ANOVA with Fisher’s post hoc test for multiple comparisons. *P < 0.05.

DISCUSSION

In this current study, we present data that indicate two major conclusions: 1) ECMR activation is a crucial mechanism whereby placental ischemia induces elevations in blood pressure and fetal growth restriction in pregnancy, and 2) leptin is a key feature of PE mediating vascular endothelial dysfunction. Collectively, in this report, we provide evidence that ECMR contributes to the PE-associated hypertension and fetal growth restriction induced by placental ischemia, while elevated leptin likely drives endothelial dysfunction in the PE milieu. The results of this study indicate that the addition of leptin administration to the RUPP model may create a more robust PE phenotype for use in the study of PE in mice, thereby enabling the widespread use of transgenic mouse models for PE.

In this study, we provide evidence indicating that late gestation placental ischemia alone, in the absence of other factors associated with PE, is not the primary mediator of vascular endothelial dysfunction in the disease. In addition, we provide the first measurement of vascular function in the RUPP mouse model of PE. It is well established that endothelial dysfunction is a hallmark of PE (34), keenly demonstrated in clinical studies by both reductions in flow-mediated dilation (10) and reduced ex vivo small artery relaxation to endothelial-dependent dilators, such as ACh (35–41). A meta-analysis by Garovic et al. (10) summarized decades of data showing that reductions in flow-mediated dilation (i.e., endothelial dysfunction) are clearly evident at the time of PE symptom onset in late gestation. In addition, their analysis demonstrated that patients who go on to develop PE characteristically develop endothelial dysfunction before the onset of symptoms, as early as 18 wk. These data indicate that endothelial dysfunction is a key component predisposing pregnant women to placental ischemia and ultimately PE in late pregnancy. Our study sheds light on a long observed, but not frequently mentioned, discrepancy of the RUPP model with the clinical presentation of endothelial dysfunction in PE. The studies of Gilbert et al. (21), Brennan et al. (22), Palei et al. (23), and Anderson et al. (24) have all reported RUPP rats do not demonstrate endothelial dysfunction in late gestation in either carotid arteries or mesenteric resistance arterioles, as assessed by endothelial-dependent relaxation to ACh. Two older studies do indicate decreases in EC₅₀ of ACh-mediated relaxation in aorta strips (42) and uterine arteries (43), however, in both studies no significant difference was found for overall relaxation (i.e., repeated-measures 2-way ANOVA) nor maximal relaxation (EC_max). Our data in this report concur with most of the literature of the RUPP rat in concluding that the placental ischemia alone induced by RUPP in mice does not confer endothelial dysfunction in late gestation in resistance arteries or thoracic aorta, as assessed by a lack of reduction of ACh-mediated relaxation. Therefore, these rodent model data indicate that endothelial dysfunction may predispose pregnant women to placental ischemia but may not be caused by placental ischemia.

A very exciting finding in our current report is that the addition of leptin infusion to the RUPP procedure in mice induces a significant reduction in endothelial-dependent relaxation to ACh in late gestation. We have previously demonstrated that leptin infusion in mid-late gestation induces vascular endothelial dysfunction in pregnant mice in the absence of placental ischemia (29). Leptin is an important reproductive hormone that is produced by adipocytes and in pregnancy is additionally produced by the placenta, leading to pregnancy-induced elevations in plasma leptin (44, 45). However, numerous clinical reports show that patients with PE present with inappropriate elevations in plasma leptin levels as early as midgestation, and that leptin levels positively correlate with adverse fetal outcomes (9, 46–59). A finding that has complicated the field of leptin in pregnancy is that preclinical reports as well as our own observations indicate that rodent placenta do not produce leptin comparable to humans in pregnancy (33). Therefore, these data indicate that the mechanism whereby patients with PE develop endothelial dysfunction may be elevations in placental leptin production, which has not been identified in rodents because of their inability to produce the hormone in placental tissue. Leptin infusion in addition to placental ischemia in RUPP rodents provides an additional feature of PE to the model in the presentation of PE-associated endothelial dysfunction.

In addition to the role of leptin in PE, this study advances our understanding of the mechanisms whereby placental ischemia leads to hypertension and vascular endothelial dysfunction. Several studies indicate that ECMR activation is a sex-specific mechanism predisposing females to hypertension. Our group demonstrated that ECMR expression is upregulated in female mice and humans and upregulated further in pregnancy (28). Functionally, we have also shown that ECMR deletion protects against leptin-mediated hypertension in female mice, both pregnant (29) and nonpregnant (28). Others have reiterated these findings; Davel et al. (25) demonstrated that in obese female mice, in which plasma leptin is elevated, deletion of the ECMR protects from vascular endothelial dysfunction. The RUPP model mimics placental ischemia which is an initiating event for hypertension and adverse fetal outcomes of PE (11, 60). In our previous study, we showed that mid-late gestation leptin infusion induces clinical characteristics of PE in endothelial dysfunction, hypertension, and fetal growth restriction (29). We show in this current study that ECMR deletion protects pregnant mice from placental ischemia-induced hypertension and fetal growth restriction. These data indicate that ECMR is a critical mediator of blood pressure control and placental blood flow, i.e., fetal nutrient supply, in pregnancy. Therefore, in PE, heightened ECMR activity may promote late-gestation disease and contribute to its increased severity.

We show in Fig. 2 that heightened vascular contractility in RUPP mice is attenuated by ECMR deletion, in association with a decrease in blood pressure in ECMR knockout mice. Therefore, the mechanism whereby ECMR promotes hypertension and fetal growth restriction in the presence of placental ischemia, but in the absence of endothelial dysfunction, in our RUPP mouse model may be due to attenuation of vascular contractility. Similar to many others studying RUPP rats, we show in this report that RUPP procedure in pregnant mice induces significant increases in vascular contractility to Phe in resistance arteries (24, 61, 62). This increase in Phe-mediated contractility was abolished by ECMR deletion in RUPP mice. These data indicated that PE is characterized by a multifaceted vascular dysfunction involving both vascular endothelial dysfunction that potentially predisposes to placental ischemia and then followed by a placental ischemia-induced vascular contractility increase and change in the activity of α1 receptor agonism. A study by Jaffe and colleagues also showed that postpartum vascular contractility and angiotensin II sensitivity are promoted following exposure to sFlt-1 in pregnancy, a model of PE symptoms in mice (63). The finding in this report that ECMR deletion protects pregnant mice from placental ischemia-induced vascular contractility in association with an alleviation in hypertension and fetal growth restriction indicates that later onset vascular contractility is a mediator of placental ischemia-induced elevations in blood pressure and reduction in fetal blood flow in late gestation (64–66). In accordance, we have shown in previous studies that MR antagonism is protective against increased vascular contractility to phenylephrine in both male and female mice when the selective antagonist eplerenone is used (31), however, not when the less-specific antagonist spironolactone is used (27). Jia et al. (26) have shown that ECMR deficiency prevents aortic stiffness and fibrosis in female mice following diet-induced obesity, therefore prevention of endothelial-derived increases in vascular resistance likely reduces sympathetic tone in the smooth muscle of females. Therefore, the mechanism whereby ECMR promotes vascular contractility, which we have showed is protective in other models of endothelial dysfunction in females (67), implies either 1) cross talk from the endothelial cells to the smooth muscle [potentially inflammatory mediated (26)] to reduce sensitivity to phenylephrine-mediated adrenergic receptor agonism or 2) altered programming of the sympathetic tone of the vascular smooth muscle, potentially due to increased vascular distensibility provided by the removal of the ECMR. Future studies in the RUPP mouse are warranted to determine if eplerenone is a potential therapeutic for the alleviation of vascular contractility and other adverse pregnancy outcomes brought on by PE.

Interestingly, and in accordance with our previous reports in both pregnant mice (29) and male mice (68), we show that leptin promotes endothelial dysfunction but does not promote vascular contractility in pregnant RUPP mice in this study. Our previous reports indicate that this reduction in Phe-mediated constriction is an NO-dependent process, as it is ablated by l-NAME, at least in male wild-type mice (68). We did not perform Phe-mediated constriction curves in the presence of l-NAME in these pregnant mice, however, we show that l-NAME does not ablate differences in ACh-mediated relaxation between ECMR^+/+ + RUPP and ECMR^+/+ + RUPP + leptin mice, indicating a non-NO mechanism for leptin-derived vascular endothelial dysfunction. Further investigation is needed to determine if leptin alters sympathetic tone or receptor expression/activity in the vascular smooth muscle of pregnant mice.

In our current study we show that eNOS phosphorylation does not decrease in RUPP mice and is not increased or decreased by ECMR deletion. Increased NO bioavailability is a crucial adaptation for healthy pregnancy (69), and is impaired in some reports of patients with PE (70, 71). In addition, our data in this report show that preincubation of resistance vessels with NOS-inhibitor l-NAME does not increase or decrease ACh endothelial-dependent dilation in resistance or conduit vessels of normal pregnant versus RUPP, nor ECMR-intact versus ECMR-deficient pregnant mice. Therefore, ECMR does not appear to regulate NO bioavailability nor NO-mediated vascular responses in pregnancy.sFlt-1 levels and PLGF levels characteristically increase and decrease, respectively, in many cohorts of patients with PE (72). Interestingly in our study, and in contrast to others’ reports in rats (73, 74), sFlt-1 and PLGF levels did not increase or decrease with RUPP in our pregnant mice. Few RUPP mouse reports currently exist in the literature, however, the association of sFlt-1 and placental ischemia seems to be less consistent in mice with some (14), but not all (15), investigators reporting RUPP-induced increases in sFlt-1 in mice. In addition, our previous report indicated that leptin infusion in pregnant mice did not induce an increase in placental mRNA expression of sFlt-1 (29). Taking these data from our current study complemented by the existing literature collectively indicates that sFlt-1 is a factor associated with placental ischemia in many cases but is not a crucial mediator of PE-associated endothelial dysfunction as RUPP rats do not demonstrate systemic endothelial dysfunction despite elevations in sFlt-1. Furthermore, reports are inconsistent regarding whether the sFlt-1 model of PE induces functional endothelial impairment. One report found dysfunction in sFlt-1 pregnant mice in thoracic aorta (75), however, others report increases in contractile responses to sFlt-1 but not endothelial-dependent relaxation (63). ECMR deletion also did not decrease sFlt-1 or increase PLGF in our current study, therefore, these are not likely the mechanism whereby ECMR confers an increase in blood pressure or fetal growth restriction in PE pregnancy. Leptin addition to RUPP in ECMR intact mice did significantly increase sFlt-1-to-PLGF ratio, which may account for the mechanism whereby leptin decreases placental weight, a common finding in patients with severe PE (76) and RUPP rats (73) and in fact the Food and Drug Administration of the United States recently approved the sFlt-1-to-PLGF ratio as a predictive clinical marker for severe features of PE. Therefore, the addition of leptin to the RUPP in mouse potentially creates a more clinically relevant and potentially robust mouse model of PE for future study.

Perspectives

The model presented in this report offers worthy preclinical findings in the understanding of the contribution of leptin and ECMR in placental ischemia and endothelial function during PE. The rising cases of PE, despite the advances in healthcare, and the lack of treatment for PE other than placental delivery creates an impetus to find mechanisms and therapies. We show that placental ischemia, an initiating event for PE, induces hypertension, fetal growth restriction, and placental insufficiency in pregnant mice but without endothelial dysfunction. Genetic deletion of ECMR abrogates placental ischemia-induced hypertension, fetal growth restriction, and placental insufficiency in placental ischemia mice. We also show that leptin supplementation at midgestation is a crucial mediator of endothelial dysfunction in the presence of placental ischemia. These data indicate that there is an important role for ECMR and leptin in promoting PE syndrome. Therefore, further studies are needed targeting ECMR as well as leptin pathways to determine whether they may be valuable therapeutic options for women with PE.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S3 can be accessed at https://doi.org/10.6084/m9.figshare.25674918.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants 1R01HL169576 and 4R00HL146948 and American Heart Association Grants CDA858380 (to J.L.F.), AHA1196923 (to D.M.), and AHA1192508 (to E.M.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.M. and J.L.F. conceived and designed research; V.A., M.E., E.M., and S.O. performed experiments; D.M., V.A., M.E., E.M., S.O., and J.L.F. analyzed data; D.M. and J.L.F. interpreted results of experiments; D.M. and J.L.F. prepared figures; D.M. and J.L.F. drafted manuscript; D.M., V.A., M.E., and J.L.F. edited and revised manuscript; D.M., V.A., M.E., E.M., S.O., and J.L.F. approved final version of manuscript.