Estrogen-Related Receptor γ Serves a Role in Blood Pressure Homeostasis During Pregnancy

Abstract

Persistent hypoxia caused by shallow trophoblast invasion and poor placental perfusion may underlie the pathophysiology of preeclampsia, a leading cause of maternal and neonatal morbidity and mortality. Previously, we found that estrogen-related receptor γ (ERRγ) serves a critical and O₂-dependent role in differentiation of human trophoblasts in culture and expression of tissue kallikrein and voltage-gated K⁺ channels. In this study, we surprisingly observed that ERRγ expression was significantly increased in placentas from preeclamptic women compared with that in gestation-matched normotensive women. To further investigate a functional role for ERRγ during pregnancy, we analyzed ERRγ-deficient mice. Maternal systolic blood pressure was significantly reduced in pregnant ERRγ^+/− females bred to ERRγ^+/− males compared with that in wild-type (WT) mice and was markedly up-regulated by treatment of WT pregnant mice with the ERRγ agonist DY131. Placentas of ERRγ^+/− mice manifested increased vascular endothelial growth factor A expression compared with that in WT mice. Notably, circulating levels of the antiangiogenic factor, soluble fms-like tyrosine kinase-1, were significantly reduced in ERRγ^+/− pregnant mice as was serum aldosterone. These effects were associated with a decrease in maternal adrenal Cyp11b1 (steroid 11β-hydroxylase) and Cyp11b2 (aldosterone synthase) expression. In contrast, adrenal Cyp11b1 and Cyp11b2 mRNA were increased in pregnant WT mice treated with DY131. Moreover, chromatin immunoprecipitation and luciferase reporter assays identified Cyp11b2 as a transcriptional target of ERRγ. Collectively, these findings reveal a potential role of ERRγ in maternal blood pressure homeostasis during pregnancy and suggest that aberrant ERRγ expression may contribute to the pathogenesis of preeclampsia.

Preeclampsia (PE), a pregnancy-induced, multisystem disease that affects 6% to 8% of pregnancies (1), is characterized by hypertension, proteinuria, and increased morbidity and mortality of mother and newborn (2). The placenta is believed to play a key role in the pathogenesis of PE, with increased expression of placental antiangiogenic factors and systemic endothelial dysfunction (3). A growing body of evidence further suggests that mitochondrial dysfunction, with oxidative stress and impaired differentiation and invasion of trophoblasts, is related to the pathogenesis of PE (4, 5). On the other hand, maternal factors can predispose to PE (6); not all pregnant women with low placental perfusion develop PE (7), and women who have had PE often develop cardiovascular diseases later in life (8). In this regard, aberrant regulation of the renin-angiotensin-aldosterone system (RAAS) may also contribute to PE (3). Despite these conjectures and the severity and consequences of this disorder, the molecular mechanisms that contribute to the pathogenesis of PE remain incompletely defined.

Estrogen-related receptor γ (ERRγ), a member of the ERR/nuclear receptor 3 subfamily of orphan nuclear receptors (9), controls gene networks involved in mitochondrial activity and biogenesis (10, 11). In studies of gene-targeted mice, ERRγ was found to regulate potassium (K⁺) homeostasis in heart, stomach, and kidney via control of a number of hypertension-associated genes (12). Whereas, ERRγ heterozygous (ERRγ^+/−) mice were apparently normal (13), ERRγ-null mice died by the second day of life, manifesting elevated serum K⁺ and cardiac arrhythmias with prolonged QT intervals (12). Interestingly, ERRγ single nucleotide polymorphisms have been correlated with altered blood pressure (BP) in humans (12), suggesting that ERRγ may affect BP homeostasis. ERRγ was found to be expressed at the highest levels in placenta among several human reproductive tissues (14). Our previous studies using human trophoblasts in primary culture revealed that ERRγ serves as a novel O₂ responsive transcription factor that is required for syncytiotrophoblast differentiation (15) and up-regulates voltage-gated K⁺ channels KCNQ1, KCNE1, and tissue kallikrein (KLK1) expression during trophoblast differentiation (16).

In this study, we unexpectedly observed increased ERRγ expression in placentas from preeclamptic women compared with those from women with normotensive pregnancies. Moreover, pregnant ERRγ^+/− female mice bred to ERRγ^+/− males had lower systolic BP, decreased circulating aldosterone, and reduced adrenal expression of Cyp11b1 and Cyp11b2, genes involved in aldosterone synthesis. Although cross-breeding experiments revealed a critical role of the maternal genotype, effects of ERRγ haploinsufficiency on BP were enhanced by pregnancy. Notably, placentas of ERRγ-deficient fetuses manifested increased expression of vascular endothelial growth factor (VEGF) A and endomucin (markers of vasculogenesis) compared with those of wild-type (WT) mice. Thus, ERRγ may serve a role in BP homeostasis and aberrant ERRγ expression may contribute to hypertensive disorders of pregnancy.

Materials and Methods

Human subjects

Term placental tissues from control normotensive women and preeclamptic women were obtained from Parkland Memorial Hospital (Dallas, Texas). Consent forms and protocols were approved by the institutional review board of the University of Texas Southwestern Medical Center. The placental tissues were washed with cold 1% NaCl immediately after delivery, stored in RNAlater (QIAGEN GmbH) solution, and snap-frozen at −80°C. PE was defined as the onset of hypertension after 20 weeks of gestation (systolic and diastolic BP ≥140 and ≥90 mm Hg, respectively, measured at 2 different time points, 4 hours to 1 week apart), along with proteinuria (≥300 mg in a 24-hour urine collection or 2 random urine specimens obtained 4 hours to 1 week apart containing ≥1+ protein by dipstick or 1 dipstick demonstrating ≥2+ protein).

Mouse model

All protocols were approved by the institutional animal care and use committee of University of Texas Southwestern Medical Center. Mice were housed in a temperature-controlled (23°C) room with a 12:12-hour light-dark cycle. ERRγ^+/− (heterozygous) mouse embryos were obtained from Mutant Mouse Regional Resource Center (University of North Carolina, Chapel Hill, North Carolina). Embryos were implanted into uteri of pseudopregnant females in the Transgenic Core, University of Texas Southwestern Medical Center. The resulting ERRγ-deficient offspring were backcrossed to WT C57BL/6 mice for at least 8 generations to establish ERRγ^+/− mice on a pure C57BL/6 background. The breeding scheme for timed pregnancy was as follows: WT females were mated with WT males; ERRγ^+/− females were mated with ERRγ^+/− males; WT females were mated with ERRγ^+/− males; and ERRγ^+/− females were mated with WT males. Female mice were housed with male mice overnight and separated the following morning. Noon of the day a vaginal plug was found was designated as 0.5 days post coitum (dpc). Maternal weight was measured at 0.5 dpc and at 9.5 to 12.5 dpc to assess pregnancy. At 18.5 dpc, pregnant mice were killed by cervical dislocation, and the uterus was removed into RNase-free, ice-cold PBS. Each implantation site was individually opened, the corresponding fetus and placenta were removed and weighed, and a fetal sample was collected for genotyping. Placentas were removed from the uterine wall, and some were microdissected into labyrinth-rich and spongiotrophoblast-rich tissues under microscopy. Kidneys and adrenal glands of mothers were removed at same time. In some cases, adrenals were decapsulated; the glomerulosa-enriched capsules were stored separately from the fasciculata-enriched remainder of the glands. Placentas, kidneys, and adrenal tissues were either flash-frozen in liquid nitrogen for RNA or protein analysis, immersion-fixed in 4% paraformaldehyde for light microscopy or cross-linked in PBS-formaldehyde (1% formaldehyde final concentration) for chromatin immunoprecipitation (ChIP) assay.

To determine the effects of ERRγ activation in vivo, WT pregnant mice were injected subcutaneously with ERRγ agonist DY131 (Tocris Biosciences), 0.8 mg/mouse dissolved in 100 μL of dimethyl sulfoxide (DMSO) (American Type Culture Collection) or with 100 μL of DMSO vehicle from 10.5 to 17.5 dpc. To determine the effect of ERRγ on BP modulation via the VEGF-endothelial nitric oxide synthase (eNOS) pathway, pregnant mice were treated for 7 days with the nitric oxide synthase inhibitor N^G-nitro-l-arginine methyl ester (l-NAME, Sigma-Aldrich) at 100 mg/L in the drinking water, beginning at 11.5 dpc. Water was refreshed every day. To determine whether sodium homeostasis was impaired, pregnant mice were treated by gavage administration of 2% sodium chloride (10 mL/kg/d) from 10.5 to 18.5 dpc.

Mouse BP measurement

Maternal arterial BP and pulse were measured in nonsedated mice using an automated tail cuff plethysmography system (BP-2000; Visitech Systems). Measurements were taken on 5 consecutive days during pregnancy from 14.5 to 18.5 dpc. Blood pressure measurements in nonpregnant mice were also taken on 5 consecutive days. BP in each mouse was measured 20 consecutive times, and the results were averaged. All BP measurements were obtained between 8:00 am and noon, before any treatment was administered.

Mouse urine and blood analysis

Mice were housed in metabolic cages (1 mouse/cage) and fed a standard diet and water ad libitum for 4 days from 14.5 to 18.5 dpc; the urine from each cage flowed into separate collection tubes over a 24-hour period and was harvested once daily. Mouse blood was collected at 18.5 dpc by submandibular puncture and allowed to clot for 2 hours at room temperature before centrifugation for 20 minutes at 3000 × g. Serum was harvested and stored at −80°C. Urine and blood constituents were assayed using Chemistry Analyzer by MicroSlide technology (Vitros 250, Ortho Clinical Diagnostics; Thermo Electron). Concentrations of soluble fms-like tyrosine kinase-1 (sFlt-1) were analyzed in serum samples using a commercially available ELISA kit (R&D Systems) as per the manufacturer's directions. Serum aldosterone concentrations were determined by ELISA (DRG Diagnostics, DRG International), following the manufacturer's instructions.

Quantitative RT-PCR (qRT-PCR)

Total RNA from tissues was extracted by an miRNeasy Mini Kit (QIAGEN). RNA was treated with DNase to remove any contaminating DNA, and 2 μg were reversed transcribed using an iScript cDNA Synthesis Kit (Bio-Rad Laboratories). Primer sets are listed in Supplemental Table 1.

For quantitative analysis of mRNA, a CFX384TM Real-Time PCR Detection System (Bio-Rad Laboratories) was used with iTaq SYBR Green Supermix with ROX (Bio-Rad Laboratories) for the detection of PCR products. The cycling conditions were 95°C for 3 minutes, followed by 40 cycles of 95°C for 10 seconds and 60°C for 30 seconds. The relative fold changes were calculated using the comparative cycle times (C_t) method with RPLP0 as a reference standard.

Immunoblot analysis

Total protein extracts were prepared from tissues using radioimmunoprecipitation assay buffer (Thermo Scientific). Equivalent amounts of protein determined by the Pierce BCA protein assay kit (Thermo Scientific) were resolved by 4% to 12% bis-Tris gel (Invitrogen) electrophoresis and blotted to Hybond-P membranes (Amersham Biosciences). The primary antibodies used were rabbit polyclonal ERRγ antibody (catalog no. ab82319; Abcam Inc) and anti-β-actin (catalog no. ab8227; Abcam Inc). Horseradish peroxidase-conjugated antirabbit IgG (GE Healthcare) was used as the secondary antibody. The membranes were developed by enhanced SuperSignal West Pico Chemiluminescent Substrate or SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific).

ChIP

ChIP assays were performed on mouse adrenal glands using a Millipore ChIP Assay Kit (Upstate [Millipore]). Cross-linking was performed using formaldehyde (1% in PBS, final) at room temperature for 10 minutes with gentle rotation. Minced tissues were pelleted in a microfuge at full speed for 1 minute at 4°C. The pellets were washed, resuspended in lysis buffer containing protease inhibitor cocktail, and sonicated. Sheared chromatin was diluted in ChIP dilution buffer and precleared following the manufacturer's instructions. For ChIP, 1% input samples were prepared. Precleared chromatin was immunoprecipitated overnight with either rabbit anti-ERRγ polyclonal antibody or rabbit IgG (Santa Cruz Biotechnology) and 30 μL of protein A agarose/salmon sperm DNA. The beads were washed following the manufacturer's instructions; cross-linking was reversed by incubation in 1% SDS and 0.1 M NaHCO₃ at 65°C overnight. DNA was then extracted from samples by phenol/chloroform (Sigma-Aldrich). Real-time quantitative PCR was performed using iTaq SYBR Green Supermix with ROX and primer pairs for the Cyp11b1 and Cyp11b2 promoter regions, listed in Supplemental Table 1. PCR was performed using the following conditions: 95°C for 3 minutes, followed by 40 cycles of 95°C for 10 seconds and 60°C for 30 seconds.

Luciferase reporter assays

The Cyp11b2-luc plasmid was generated by subcloning a fragment composed of −718 bp of sequence upstream and +96 bp downstream of the transcription start site of the mouse Cyp11b2 gene PCR amplified from mouse genomic DNA and cloned into a pGL4.12 vector (Promega). A putative ERRγ response element (ERRE) (5′AAGGTC) site identified by the Genomatix MatInspector program (http://www.genomatix.de/) located at −242 bp was mutated to (5′AGATTC) using a QuikChange II Site-Directed Mutagenesis Kit (Stratagene) to create the Cyp11b2m-luc plasmid. For luciferase reporter assays, human embryonic kidney (HEK) 293 cells were seeded in 24-well plates and transfected using FuGENE HD transfection reagent (Roche) with WT or mutated Cyp11b2 reporter constructs, mouse ERRγ expression vectors, and Renilla luciferase plasmid. Forty-eight hours after transfection, cells were harvested in Passive Lysis Buffer (Promega). Firefly luciferase and Renilla luciferase activities were assayed using a Dual-Luciferase Assay System (Promega). Relative luciferase activities were calculated by normalizing firefly luciferase activity to Renilla luciferase activity in the same samples to correct for transfection efficiencies.

Immunohistochemical studies

Mouse placentas were fixed in 4% paraformaldehyde, embedded in paraffin, and cut into 5-μm sections. Sections were deparaffinized and stained with hematoxylin-eosin (H&E). For immunohistochemical staining, sections were deparaffinized and incubated with 5% normal rabbit serum (Vector Laboratories Inc) for 30 minutes to prevent nonspecific staining. The primary antibody was rat monoclonal anti-endomucin (V.7C7) antibody (1:1250, catalog no. sc-65495; Santa Cruz Biotechnology); the secondary antibody was rabbit antirat IgG (1:500; Vector Laboratories Inc). PBS was added in place of the primary antibody as a negative control. Nuclei were counterstained with hematoxylin. All procedures were carried out according to the manufacturer's instructions.

For immunohistochemical analysis of β-galactosidase in mouse adrenals, WT and ERRγ^+/− (LacZ was knocked into the ERRγ locus in construction of the knockout) adult nonpregnant female mice were transcardially perfusion fixed with 4% paraformaldehyde in PBS (4% PFA). Intact kidneys and adrenals from each genotype, complete with capsule, were dissected and further fixed by immersion in 20 vol of 4% PFA for 16 hours at 4°C with constant agitation. Tissues were briefly rinsed, dehydrated, cleared, and paraffin embedded by standard procedures. Paraffin blocks were sectioned at 5 μm, dewaxed, and subjected to antigen retrieval at 95°C in 1× Antigen Retrieval Citra solution (BioGenex), blocked with 0.5% blocking reagent (supplied with a Tyramide Signal Amplification kit, catalog no. NEL749A; PerkinElmer), and incubated with rabbit polyclonal antibody to β-galactosidase (1:500, catalog A11132; Invitrogen Life Technologies) overnight at 4°C. Bound primary antibody was detected with biotinylated goat antirabbit IgG (1:200; Vector Laboratories) and subsequently amplified using Tyramide Signal Amplification according to the manufacturer's instructions. Detection was finalized with DAB Chromagen (Dako) before hematoxylin counterstaining, dehydration, clearing, and coverslipping.

Data analyses

Data are expressed as means ± SEM. Differences between groups were analyzed by multifactorial ANOVA or the Student t test with SPSS 16.0 software. Statistical significance was set as P < .05.

Results

ERRγ expression is increased in placentas from preeclamptic women

Previous studies in ERRγ knockout mice suggested its role in K⁺ homeostasis in heart, stomach, and kidney (12). Moreover, using human trophoblasts in primary culture, we recently observed that ERRγ served a crucial, O₂-dependent role in syncytiotrophoblast differentiation (15) and in the expression of voltage-gated K⁺ (K_V7) channels and KLK1 (16). To gain further insight into the possible role of placental ERRγ during pregnancy, we analyzed ERRγ expression in term placentas from 9 preeclamptic and 8 gestation-matched normotensive women by qRT-PCR and immunoblotting. The clinical characteristics of these subjects, listed in Supplemental Table 2, indicate that the preeclamptic women had markedly elevated systolic and diastolic BP and proteinuria compared with those for the normotensive group. Remarkably, both ERRγ mRNA and protein expression were significantly increased in placentas from preeclamptic women compared with those from control women (Figure 1, A–C).

ERRγ^+/− mice have lower BP

Because ERRγ-null mice die during the first 72 hours of life, we investigated the phenotype of ERRγ^+/− pregnant mice that had been bred to genetically like males. The ERRγ^+/− females were compared with WT females mated to WT males as controls. BP was measured daily from 14.5 to 18.5 dpc using a tail cuff method. Notably, the ERRγ^+/− pregnant mice had significantly lower BP than the WT mice (Figure 2A). Because ERRγ haploinsufficiency caused maternal hypotension, we investigated whether increasing ERRγ activity by treatment of WT pregnant mice with the ERRγ agonist DY131 could induce hypertension. The pregnant mice treated with DY131 developed significant hypertension from 16.5 dpc (Figure 2B); however, the urine albumin to creatinine ratio, commonly found to be elevated in PE, was unaltered by DY131 treatment (Supplemental Figure 1A). DY131 also is an agonist for ERRβ (17); however, it is likely that the observed effects of DY131 on maternal BP homeostasis were manifest through ERRγ, because ERRβ is involved in early placental development and its expression is not detected in placenta after 8.5 dpc (18).

To determine whether the effect of ERRγ on BP was due to a maternal haploinsufficiency or fetal/placental deficiency in ERRγ expression, we mated WT females with ERRγ^+/− males and ERRγ^+/− females with WT males. Despite having fetuses with similar genotype compositions, ERRγ^+/− pregnant females bred to WT males had significantly lower BP than WT females bred to ERRγ^+/− males from 15.5 to 18.5 dpc (Figure 2C). Moreover, the BP of WT females mice mated either to WT or to ERRγ^+/− males did not differ from one another (Supplemental Figure 2A, data extracted from Figure 2). On the other hand, the BP of ERRγ^+/− females mated with ERRγ^+/− males was marginally but significantly lower than that of ERRγ^+/− females mated with WT males (Supplemental Figure 2B, data extracted from Figure 2). This finding suggests that the presence of ERRγ^−/− fetuses, although comprising as little as 25% of the litter, can affect BP during pregnancy. Collectively, these results suggest that both maternal haploinsufficiency and fetal/placental deficiency of ERRγ contribute to the hypotensive phenotype. However, because ERRγ^−/− pups comprised only ∼25% of the total litter of ERRγ^+/− females mated with ERRγ^+/− males, the fetal/placental contribution of ERRγ deficiency on BP may be underestimated.

We then investigated whether the hypotension observed in ERRγ^+/− female mice was pregnancy induced. We measured the BP of nonpregnant ERRγ^+/− and WT female mice for 5 consecutive days and found that nonpregnant ERRγ^+/− female mice also had lower BP than nonpregnant WT females; however, the difference between the two groups was not as pronounced as that during pregnancy (Figure 2D). This result indicates that increased blood flow during pregnancy may augment the effects of ERRγ deficiency on the regulation of BP. We also measured BP of ERRγ^+/− and WT female mice at 9 to 13 days postpartum and found that although BP increased in both groups after delivery, the ERRγ^+/− mice maintained lower BP than WT females (Supplemental Figure 2C).

ERRγ deficiency has proangiogenic effects in the placenta

Because the placenta is thought to contribute to the pathogenesis of PE, we initially analyzed morphological differences in placentas of WT and ERRγ-deficient mice. ERRγ mRNA was expressed both in the labyrinth and spongiotrophoblast regions of WT mouse placenta, and ERRγ protein and mRNA were confirmed to be absent in ERRγ^−/− placentas (Supplemental Figure 3). Placental weights of ERRγ^+/− pregnant mice were significantly increased compared with those from WT mice (Supplemental Table 3), and fetuses from ERRγ^+/− females mated with ERRγ^+/− males were smaller than those from WT females mated with WT males (Supplemental Table 3). Neither weights of the fetuses nor weights of the placentas were different among ERRγ^−/−, ERRγ^+/−, or WT littermates from ERRγ^+/− females mated with ERRγ^+/− males (Supplemental Table 3). Consistent with increased placental weight, H&E staining indicated that ERRγ^−/− placentas from these ERRγ^+/− dams contained increased numbers of blood cells (Figure 3A, left panels, a vs b) and immunohistochemical staining showed that endomucin, a marker of endothelial cells, was expressed at higher levels in the labyrinths of ERRγ^−/− placentas from ERRγ^+/− dams than in WT placentas from WT dams (Figure 3A, right panels, c vs d; panel e is the negative control). Moreover, endomucin mRNA levels were significantly increased in placentas of ERRγ^−/− fetuses compared with those in WT fetuses (Figure 3B). We next investigated the expression of VEGFA, one of the major angiogenic factors in placenta. In association with the absence of ERRγ mRNA and protein (Supplemental Figure 3), VEGFA mRNA levels were significantly increased in placentas of ERRγ^−/− fetuses compared with those in WT placentas from WT females bred to WT males (Figure 3C). Notably, levels of the antiangiogenic factor sFlt-1 were significantly decreased in serum of ERRγ^+/− compared with that of WT pregnant mice (Figure 3D). Because sFlt-1 is primarily secreted by the placenta (19), its circulating levels were markedly reduced in the nonpregnant mice compared with that in pregnant mice and did not differ between ERRγ^+/− and WT nonpregnant mice (Figure 3D). sFlt-1 inhibits VEGF signaling to activate eNOS, promote vasodilation, and affect BP (20).

To determine whether ERRγ deficiency affects BP through the VEGF-eNOS pathway, we treated ERRγ^+/− and WT pregnant mice bred to ERRγ^+/− and WT males, respectively, with the eNOS inhibitor l-NAME in their drinking water (100 mg/L) from 11.5 to 18.5 dpc. The BP of both WT and ERRγ^+/− pregnant mice was significantly increased after l-NAME treatment compared with that of the untreated controls (Supplemental Figure 2, D and E). However, the l-NAME–treated ERRγ^+/− pregnant mice had significantly lower BP than that of l-NAME–treated WT pregnant mice (Figure 3E). Thus, the differences in BP between WT and ERRγ^+/− mice were maintained in the presence of l-NAME, and BP was increased equivalently in both groups.

Although our collective findings suggest that ERRγ haploinsufficiency has a proangiogenic effect in the placenta, there appear to be additional maternal effects that contribute to the hypotensive phenotype of these mice. Therefore, we conducted studies to assess the role of ERRγ on salt homeostasis and on the maternal RAAS.

ERRγ^+/− pregnant mice have hyponatremia and increased salt wasting

Because prior studies with ERRγ knockout mice indicated that ERRγ regulated cardiac, gastric, and renal K⁺ homeostasis (12), we compared renal function and electrolyte handling in pregnant ERRγ^+/− females mated with ERRγ^+/− males vs WT females mated with WT males. Interestingly, ERRγ^+/− pregnant females excreted more Na⁺ (Figure 4A) and more K⁺ (Supplemental Figure 1B) than the WT pregnant mice. Serum Na⁺ concentrations were significantly decreased in ERRγ^+/− pregnant females compared with those in WT females (Figure 4B), which probably contributed to the hypotension. However, serum concentrations of K⁺ in ERRγ^+/− pregnant females were not different from those in WT females (Supplemental Figure 1C). ERRγ deficiency was previously shown to reduce expression of voltage-gated K⁺ channels in the kidney (12). We, therefore, suggest that decreased K⁺ channel expression may negatively affect K⁺ reabsorption by the renal tubules and contribute, in part, to the absence of an apparent change in serum K⁺ levels. When ERRγ^+/− and WT pregnant mice were treated with high salt administered by gavage, both groups manifested increased serum Na⁺ concentrations (Figure 4B). Notably, the high-salt treatment eliminated the difference in serum Na⁺ concentrations (Figure 4B) and in total urine Na⁺ (Figure 4C) between the two groups. Consequently, the difference in BP between ERRγ^+/− and WT mice was no longer evident after high-salt treatment (Figure 4D). Surprisingly, BP in WT pregnant mice was significantly decreased by high-salt treatment (Supplemental Figure 2F), whereas BP was slightly but not significantly increased in ERRγ^+/− pregnant mice treated with high salt (Supplemental Figure 2G). High-salt treatment, therefore, eliminated the difference in BP between WT and ERRγ^+/− mice (Figure 4D). These combined results suggest that altered Na⁺ and K⁺ handling in ERRγ^+/− pregnant mice contributes to maternal hypotension.

ERRγ deficiency decreases aldosterone production by reducing Cyp11b1 and Cyp11b2 expression

Given the important role of the RAAS on Na⁺ and K⁺ homeostasis and BP regulation, we analyzed expression of RAAS components in ERRγ^+/− and WT pregnant mice. mRNA expression of several members of the RAAS, including angiotensinogen, renin, angiotensin I converting-enzyme 1 (Ace1), angiotensin I-converting enzyme 2 (Ace2) and angiotensin II type 1a receptor (Agtr1a) were found to be similar in the kidneys of ERRγ^+/− vs WT pregnant mice (Supplemental Figure 4). Interestingly, serum aldosterone levels were significantly increased in WT pregnant compared with those in nonpregnant mice and were significantly decreased in hypotensive ERRγ^+/− pregnant females to levels similar to those of WT and ERRγ^+/− nonpregnant animals (Figure 5A). Thus, ERRγ haploinsufficiency prevented the pregnancy-induced increase in circulating aldosterone levels observed in WT mice (Figure 5A) and had a less pronounced effect on BP (Figure 2D) than in pregnant mice. This suggests a role of pregnancy in manifesting the phenotype.

Aldosterone is synthesized from cholesterol within the zona glomerulosa of the adrenal cortex by the enzymes, cholesterol side-chain cleavage enzyme (Cyp11a1), 3β-hydroxysteroid dehydrogenase (Hsd3b2) type 2, steroid 21-hydroxylase (Cyp21), steroid 11β-hydroxylase (Cyp11b1), and aldosterone synthase (Cyp11b2). We observed that expression of Cyp11a1 and Cyp21 in the maternal adrenals was unaffected by genotype (Supplemental Figure 5); however, in accord with reduced ERRγ mRNA levels in the maternal adrenals (Figure 5B), Cyp11b1 and Cyp11b2 mRNA levels were significantly decreased in the adrenals of pregnant and nonpregnant ERRγ^+/− females compared with those in WT females (Figure 5, C and D). Interestingly, the levels of ERRγ and Cyp11b2 mRNA were increased in nonpregnant compared with those in pregnant mice (Figure 5, B and D), which correlates with the significantly higher BP in the nonpregnant animals (Figure 2D). Furthermore, treatment of pregnant WT mice with the ERRγ agonist DY131 caused a significant increase in Cyp11b1 (Figure 5E) and Cyp11b2 (Figure 5F) mRNA expression. Cyp11b2 has been reported to be expressed exclusively in the zona glomerulosa of mouse adrenal, whereas Cyp11b1 is expressed in both the zona fasciculata and zona glomerulosa (21). To assess the localization of ERRγ in mouse adrenals, we used a two-pronged approach. In the first, we removed the capsule (which contains adherent glomerulosa cells) from adrenal glands of adult WT and ERRγ^+/− nonpregnant female mice and compared ERRγ and Cyp11b2 mRNA expression in the capsule-adhering cells and in the remaining gland (enriched in fasciculata). As can be seen in Figure 5G, both ERRγ and Cyp11b2 mRNA were significantly increased in the glomerulosa-enriched capsule-adhering cells compared with those in the remaining gland; both were significantly increased in WT compared with that in ERRγ^+/− adrenals. In the second approach, we took advantage of the fact that the LacZ gene was knocked into the ERRγ locus in creating the knockout mice (13) and used immunohistochemistry to analyze β-galactosidase protein in adrenals of adult nonpregnant ERRγ^+/− females vs WT females as a negative control (Figure 5H). As can be seen by the brown staining, β-galactosidase protein was expressed at higher levels in the adrenal zona glomerulosa than in the zona fasciculata of the ERRγ^+/− adrenals (Figure 5H). These combined findings suggest that ERRγ is enriched in the zona glomerulosa of mouse adrenals where it regulates expression of Cyp11b2 and Cyp11b1.

ERRγ regulates Cyp11b2 promoter activity

In consideration of the singular role of Cyp11b2 in aldosterone synthesis, it was of interest to determine whether the Cyp11b2 gene promoter is directly regulated by binding of ERRγ. We, therefore, conducted ChIP analysis using adrenal glands from nonpregnant ERRγ^+/− and WT female mice. A putative ERRE identified within the Cyp11b2 5′-flanking region using the Genomatix MatInspector program was found to specifically bind endogenous ERRγ (Figure 6A). Furthermore, ERRγ binding to the Cyp11b2 promoter region was significantly reduced in adrenals of ERRγ^+/− mice (Figure 6B). Notably, ERRγ binding in the ERRγ^+/− adrenals was undetectable over that of the nonimmune IgG control, suggesting that ERRγ levels in adrenals of ERRγ^+/− mice may be below the threshold for binding. Moreover, in cotransfection assays of HEK293 cells with a fusion gene composed of a luciferase reporter fused downstream of −347 bp of the Cyp11b2 5′-flanking sequence, cotransfection of the ERRγ expression vector caused a 2.2-fold induction of luciferase activity compared with cotransfection of an empty expression vector (Figure 6B). This induction was prevented when the putative ERRE was mutated (Figure 6B). Taken altogether, these findings suggest that ERRγ haploinsufficiency causes reduced aldosterone synthesis, resulting in poor Na⁺ and K⁺ handling and maternal hypotension.

Discussion

ERRγ, a constitutively active orphan nuclear receptor, is highly expressed in placenta (14). In recent studies of cultured human trophoblasts, we observed that ERRγ up-regulated expression of K_V7 channels and tissue KLK1 in an O₂-dependent manner (16). Based on the central role of placental hypoxia in the pathogenesis of PE, in the present study, we analyzed ERRγ expression in placentas of women with PE vs gestation-matched normotensive subjects. Surprisingly, we observed that ERRγ mRNA and protein levels were significantly increased in preeclamptic placentas. Based on these intriguing findings, we characterized ERRγ gene targeted mice. Because ERRγ-null fetuses die at birth, we initially studied ERRγ^+/− females bred to genetically like males. As summarized in Figure 6C, we observed that ERRγ haploinsufficiency had effects on both the maternal adrenals and the placenta. Notably, ERRγ^+/− pregnant mice manifested hypotension associated with decreased circulating aldosterone and reduced renal Na⁺ reabsorption. The decreased Na⁺ reabsorption in the ERRγ^+/− mice was probably caused by decreased expression of Cyp11b1 and Cyp11b2 and reduced aldosterone production by the maternal adrenals. In contrast, treatment of WT pregnant mice with the ERRγ agonist DY131 resulted in increased systolic BP, coupled with increased Cyp11b1 and Cyp11b2 expression in the maternal adrenals. Notably, Cyp11b2-deficient (AS^−/−) mice were reported to have reduced systolic BP that was further decreased during pregnancy (22). In the present study, ChIP assays of mouse adrenal tissues and luciferase reporter assays revealed Cyp11b2 as a transcriptional target of ERRγ.

Aldosterone, the key hormone in Na⁺ and K⁺ homeostasis, fluid balance, and BP exerts its effects in the distal nephron by promoting reabsorption of Na⁺ and excretion of K⁺, resulting in volume expansion (23, 24). Biosynthesis of aldosterone from cholesterol occurs through a series of enzymatic steps that culminate in CYP11B1- and CYP11B2-catalyzed reactions within the adrenal zona fasciculata and zona glomerulosa (25). Zone-specific expression of CYP11B2 contributes to glomerulosa-specific aldosterone production (21). We observed that ERRγ was enriched in the zona glomerulosa of mouse adrenal, where Cyp11b2 is exclusively expressed.

During normal pregnancy in women, aldosterone levels increase markedly (26), due primarily to RAAS up-regulation. However, most pregnant women are normotensive, presumably due to increased placental production of progesterone, which antagonizes aldosterone action (27) and maintains BP homeostasis. If this balance is perturbed, alterations in BP can occur (22). Remarkably, although renal blood flow and glomerular filtration rate increase markedly during pregnancy, this expansion of plasma volume and the decrease in vascular resistance are less pronounced in preeclampsia. Consistent with these changes, aldosterone is reduced in preeclamptic vs normotensive pregnancies (28, 29). However, it is unclear whether the apparent down-regulation of the RAAS in preeclampsia is a physiological response to the increased BP or a contributor to the pathophysiology of the disease (28). Our present findings revealed that although circulating aldosterone levels were significantly increased in pregnant WT compared with those in ERRγ^+/− mice, aldosterone levels did not differ between nonpregnant ERRγ^+/− and WT females. Likewise, it was reported that aldosterone and Na⁺ levels were comparable between ERRγ null and WT pups (12). This finding suggests that some aspect of pregnancy (possibly the impact of an ∼50% increase in blood volume and in cardiac output) enhances aldosterone production in the WT mothers, whereas the ERRγ haploinsufficient mothers are unable to respond.

We found that ERRγ was expressed in the highly vascularized placental labyrinth and that the endothelial markers, endomucin and VEGFA, were increased in ERRγ-null placentas compared with those in WT placentas. In contrast, Matsakas et al (30) reported that ERRγ has a proangiogenic effect in skeletal muscle and demonstrated ERRγ binding to conserved regions within the VEGFA promoter region. On the other hand, VEGF protein was found to be significantly increased in placentas of sheep with low aldosterone levels (31), whereas aldosterone blockade by spironolactone in the developing rat kidney up-regulated renal VEGF expression (32). Thus, ERRγ deficiency may enhance VEGF expression in placenta, in part, by reducing aldosterone synthesis by the maternal adrenals.

Based on our previous studies with cultured human trophoblasts (15, 16) and the present compelling findings, we propose that ERRγ might serve a gestation-dependent, biphasic role during pregnancy. We suggest that early in gestation when the placenta is relatively hypoxic, ERRγ expression is maintained at relatively low levels (15). The hypoxia and decreased ERRγ expression promote cytotrophoblast proliferation and remodeling of the spiral arteries (33, 34), increased VEGFA expression, and enhanced angiogenesis required for adequate placental perfusion. As a result, placental O₂ tension and ERRγ expression increase to promote trophoblast differentiation and production of placental hormones required for the maintenance of pregnancy. The increased placental ERRγ also stimulates expression of K_V7 channels and KLK1, which maintain placental vascular tone. Notably, ERRγ also enhances placental expression of 11β-hydroxysteroid dehydrogenase type 2 (16) to increase cortisol metabolism and protect the developing fetus and placenta from potentially hypertensive effects of elevated maternal glucocorticoids (35). Within the maternal compartment, ERRγ may modulate BP homeostasis by regulating production of aldosterone by the adrenal zona glomerulosa. On the other hand, based on the effects of ERRγ in suppressing placental VEGFA expression and its antiproliferative effects in breast cancer cells (36), we suggest that aberrant up-regulation of ERRγ in early pregnancy may contribute to the pathogenesis of PE by inhibiting cytotrophoblast proliferation and invasion, as well as vasculogenesis. This would promote a prolonged hypoxic state with enhanced sFlt-1 expression, further compromising placental perfusion and causing widespread endothelial dysfunction within the maternal circulation (37). The abnormal up-regulation of ERRγ may enhance aldosterone production by the maternal adrenals, further contributing to the hypertensive phenotype. In the future, mouse models should be used to more directly assess the consequences of placenta-specific ERRγ dysregulation on BP homeostasis during pregnancy.

Nuclear receptors are among the most effective therapeutic targets for a multitude of disorders, including hypertension, cancer, cardiovascular disease, and metabolic syndrome (38). Although a natural ligand(s) for ERRγ remains to be discovered, a number of synthetic ligands have been identified that alter ERRγ functional activity (39, 40). In light of our present findings, we suggest that ERRγ may provide a promising therapeutic target to reduce hypertension in preeclamptic women.

Acknowledgments

We acknowledge the postdoctoral trainee exchange program between the University of Texas Southwestern and the First Affiliated Hospital of Sun Yat-sen University.

This work was supported by National Institutes of Health Grant 5 R01 DK031206 (to C.R.M.). The University of Texas O'Brien Kidney Core Research Center is supported by National Institutes of Health Grant P30 DK079328.

Y.L, P.K., and C.R.M. designed the research. Y.L., P.K., C-C.C., J.L., L.W., J.M.S., and L.M. performed research. Y.L., P.K., and C.R.M. analyzed data. C.T. and J.M.A provided reagents. Y.L., P.K. and C.R.M. wrote the manuscript.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• BP

  blood pressure

 
• ChIP

  chromatin immunoprecipitation

 
• DMSO

  dimethyl sulfoxide

 
• dpc

  days post coitum

 
• eNOS

  endothelial nitric oxide synthase

 
• ERRγ

  estrogen-related receptor γ

 
• ERRE

  ERRγ response element

 
• H&E

  hematoxylin and eosin

 
• HEK

  human embryonic kidney

 
• KLK1

  kallikrein

 
• l-NAME

  N-nitro-l-arginine methyl ester

 
• PE

  preeclampsia

 
• qRT-PCR

  quantitative RT-PCR

 
• RAAS

  renin-angiotensin-aldosterone system

 
• sFlt-1

  soluble fms-like tyrosine kinase-1

 
• VEGF

  vascular endothelial growth factor

 
• WT

  wild-type.

References