Atrial Natriuretic Peptide Regulates Zona Glomerulosa Growth and Aldosterone Release

Abstract

BACKGROUND:

The cardiac hormone ANP (atrial natriuretic peptide) maintains intravascular volume homeostasis through coordinated endothelial and renal effects. Since in vitro studies suggested that this circuit involves the inhibition of adrenal aldosterone release, here we investigated whether ANP indeed mediates endocrine heart-to-adrenal communication in vivo.

METHODS:

Mouse models with zona glomerulosa (ZG)-restricted deletion of the ANP receptor, GC-A (guanylyl cyclase-A), or the downstream cyclic GMP-stimulated PDE (phosphodiesterase) 2A were established through Cre/LoxP technology.

RESULTS:

Immunoblots and immunohistochemistry demonstrated high expression of GC-A and PDE2A in ZG cells. Cell-specific deletion of GC-A led to mildly expanded ZG, with more aldosterone synthase-expressing cells and elevated plasma aldosterone. While in control mice a high-salt diet reduced ZG thickness and aldosterone production, in their ZG GC-A knockout littermates these responses were blunted, provoking hypervolemic hypertension. Notably, ZG PDE2A knockout mice did not exhibit such alterations. In cultured human adrenal cells, ANP counteracted the stimulatory effects of Angiotensin II on aldosterone synthase expression and cell proliferation pathways through activation of cGMP-dependent protein kinase I. Notably, in human aldosterone-producing adrenal adenomas, the expression of GC-A was markedly lower than in healthy adrenal cortical tissue, suggesting reduced ANP/GC-A signaling in the tumor cells.

CONCLUSIONS:

ANP-mediated heart-adrenal communication moderates ZG cell growth and aldosterone release and participates in the inhibitory effects of dietary salt/volume load. Conversely, impaired ANP/cGMP signaling leads to excessive ZG cell growth and aldosterone production. Reinforcing this pathway with novel antibody- or peptide-based therapies may benefit patients with hyperaldosteronism and resistant hypertension.

Novelty and Relevance.

What Is New?

The ANP (atrial natriuretic peptide) receptor, GC-A (guanylyl cyclase A), and its downstream cGMP-stimulated messengers, PDE (phosphodiesterase) 2A and protein kinase (cGK) I, are expressed in human and murine adrenal zona glomerulosa (ZG) cells.

Our studies in genetic mouse models demonstrate that endocrine ANP, via GC-A/cGMP signaling, moderates ZG cell growth and aldosterone release, preventing increases in intravascular volume and blood pressure by dietary salt load.

This role is independent of PDE2A and probably mediated by cGKI (cGMP-dependent protein kinase I).

What Is Relevant?

GC-A expression is diminished in human aldosterone-producing adrenal adenomas.

Clinical/Pathophysiological implications?

Impaired ZG regulation by ANP may participate in enhanced growth and aldosterone production of Conn adenomas.

Understanding the role of ANP in aldosterone regulation may define novel principles in the treatment of primary aldosteronism and resistant hypertension.

The adrenal hormone aldosterone contributes to maintaining physiological body salt and water homeostasis and arterial blood pressure (ABP). Dysregulated, increased aldosterone secretion facilitates sodium retention, hypokalemia, and arterial hypertension.¹ Primary aldosteronism (PA), characterized by renin-independent aldosterone production, is the most common endocrine cause of secondary, often therapy-resistant hypertension, with an overall prevalence between 5 and 13% in hypertensive patients.¹ Compared with essential hypertension with unaltered aldosterone, hypertension with elevated aldosterone levels causes more end-organ damage and cardiovascular morbidity, including chronic heart failure and stroke.^1,2 Moreover, large population studies recently revealed that milder and subclinical forms of PA are highly prevalent and linked to arterial stiffness, cardiac hypertrophy, and higher odds of developing hypertension.³ Even patients with only mildly increased aldosterone and normal ABP have enhanced cardiovascular risk, which is related to profibrotic and proinflammatory aldosterone actions.³ Accordingly, treatment with MR (mineralocorticoid receptor) antagonists decreases the morbidity and mortality of patients with heart and kidney diseases. Because MR-independent effects of aldosterone contribute to organ damage, aldosterone synthase (CYP11B2) inhibitors are presently being evaluated for the treatment of resistant hypertension, showing promising results.⁴ A better understanding of the pathways regulating aldosterone secretion can foster the development of aldosterone-moderating drugs for cardiovascular and renal protection.

During body salt/volume depletion, the renin-Ang II (angiotensin II) system is the main stimulus for aldosterone secretion. Ang II increases aldosterone production immediately (within minutes) through increased phosphorylation of StAR (steroidogenic acute regulatory) protein, and chronically (over hours to days) by inducing the growth of zona glomerulosa (ZG) cells and the expression of proteins involved in aldosterone synthesis, particularly CYP11B2.⁵ In contrast, increased salt/volume intake normally represses Renin-Ang II levels and Ang II-dependent ZG activity. Whether other hormones contribute to such aldosterone suppression is unknown.

The cardiac hormone ANP (atrial natriuretic peptide) counter-regulates increases in intravascular volume and ABP.⁶ Even common variants in the human ANP gene (NPPA) and small changes in ANP plasma levels contribute to interindividual ABP variations and hypertension.⁷ Via its GC-A (guanylyl cyclase-A) receptor and cyclic GMP signaling, ANP coordinates endothelial permeability and renal salt/water excretion to prevent or reverse hypervolemia.^6,8 Studies with exogenous synthetic ANP suggested that the natriuretic actions involve the inhibition of aldosterone release. Hence, very low ANP concentrations (IC 50’s between 0.1 and 0.2 nmol/L) inhibited the acute stimulatory effects of Ang II, adrenocorticotropic hormone (ACTH), or hyperkaliemia on aldosterone release from cultured bovine and rat ZG cells.^9,10 In rats, ANP infusions prevented increases in aldosterone levels in response to sodium depletion or ACTH administration, without changing plasma renin and corticosterone concentrations.^10,11

In cultured bovine ZG cells, the inhibitory effect of ANP on acute ACTH-induced aldosterone release was mediated by the cGMP-stimulated dual PDE (phosphodiesterase) 2A. Rapid PDE2A-mediated cAMP hydrolysis abolished ACTH/cAMP-dependent signaling.^12,13 Whether PDE2A also mediates inhibitory effects of ANP on ACTH-independent aldosterone production is unknown. In view of the subordinate role of the stress hormone ACTH in the regulation of ABP, this question has physiological and clinical relevance.

To dissect whether endogenous, endocrine cardiac ANP participates in the adjustment of aldosterone production to dietary salt/volume load, we generated 2 novel genetic mouse models with conditional, ZG-specific ablation of either GC-A or PDE2A. The phenotype of these mice demonstrates that cardiac ANP chronically moderates ZG growth and aldosterone production. Such effects are mediated by GC-A/cGMP signaling but independent of PDE2A. The pathophysiological implication is indicated by the observation that GC-A expression is downregulated in human aldosterone-producing adrenal adenomas.

Methods

Data Availability

The authors declare that all supporting data are available within the article and its Supplemental Material. The supplemental Major Resources Table describes primers and probes used for quantitative RT-PCR and antibodies used for immunoblotting and immunohistochemistry.

Generation of Mice With Cell-Restricted Deletion of GC-A or PDE2A in ZG Cells

All animal studies were approved by the Animal Care and Use Committee of Würzburg University (Regierung von Unterfranken, Approval number 55.2 2532-2-1772). They conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (publication no. 85-23, revised 2011) and the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes. The protocol was conducted in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines.

Cell-specific inactivation (KO) of GC-A (gene name Npr1) in ZG cells was achieved by intercrossing GC-A^{flox/flox 8} or Pde2a^flox/flox mice with AS^+/Cre mice. The AS^+/Cre (Cyp11b2-Cre) mice were kindly provided by Professor David T. Breault (Division of Endocrinology, Boston Children´s Hospital) and were previously used to delete various floxed genes in ZG cells.^14,15 For generating Pde2a^flox/flox mice, an embryonic stem cell line containing a knockout-first Pde2a allele (Pde2a^{tm1a(EUCOMM)Wtsi}) was obtained from the European Conditional Mouse Mutagenesis Program. After germline transmission, the LacZ-neo cassette was removed by crossing with a FLPe deleter strain followed by mating to C57BL/6J for at least 5 generations (scheme in Figure S1). Breeding pairs of GC-A^flox/flox or Pde2a^flox/flox male and AS^+/Cre female mice were used to generate mice with ZG-restricted deletion of GC-A or PDE2A. GC-A^flox/fox and PDE2A^flox/fox littermates without the Cre^tg served as controls within each experiment. Studies were performed with 2 to 3-month-old mice of C57Bl6/J background maintained on a normal-salt diet (NSD; 0.2% NaCl; E15430-04 from Ssniff GmbH, Soest, Germany) or subjected to a high-salt diet (HSD) for 14 days (4% NaCl; E15431-34). Water was provided at libitum. Male mice were used to minimize sex hormone-related variability and ensure reproducible phenotypes.

Studies of ABP and LV Function

ABP was measured in awake, trained mice by tail-cuff plethysmography during 5 consecutive days^8,16 between 8 and 12 am (the animal facility’s lighting schedule is from 6 am to 6 pm). On each day, after a 5-minute rest period, ABP was measured 5× within 10 minutes, and the values were averaged. The first 2 days were considered training days. The mean ABP of the next 3 days was recorded and averaged. Terminal closed-chest hemodynamic studies of left ventricular (LV) function were performed under 2% isoflurane anesthesia with a retrogradely inserted Micro-Tip pressure-volume catheter.¹⁷ Buprenorphine (0.05–0.1 mg/kg BW SC) was used for preemptive analgesia. After 20 minutes of LV recordings, the mice were euthanized by exsanguination under deep anesthesia. Atria and LV samples, as well as right adrenals and kidneys, were snap-frozen in liquid nitrogen (for protein and mRNA extraction); left adrenals and LV slices were fixed in 4% PFA for histology and immunohistochemistry. Morphometric analyses of LV Periodic Acid-Schiff-stained cardiomyocytes and Sirius red-stained interstitial collagen were performed as described.^16,17

Immunostaining and Morphometric Analyses of Mouse Adrenals

Paraffin-embedded 5 µm-thick adrenal sections were stained with antibodies against PDE2A, Disabled homolog (Dab)2, CYP11B2 (Aldosterone synthase), and DAPI (for cell nuclei). Rabbit anti-CYP11B2 antibody was kindly provided by Professor Celso E. Gomez-Sanchez at the Medical Center of the University of Mississippi. Images were acquired using an Olympus BX53 fluorescence microscope with a 20× objective. Cross-sectional PDE2A⁺ and Dab2⁺-stained ZG areas, their outer lengths (for normalization of each area), and the number of Dab2⁺/DAPI⁺ cells costained for CYP11B2 (in percent of the total counted Dab2⁺ cells) were quantified in 3 randomly selected fields from each adrenal using cellSens Dimension software (version 4.2.1; Evident, Hamburg, Germany).

Determination of Plasma ANP, Angiotensin II, Aldosterone, and Electrolyte Levels

For plasma sampling, mice were anesthetized with 2% isoflurane followed by thoracotomy. Blood was obtained between 8 and 12 am by cardiac puncture, collected into lithium heparin- or EDTA-coated tubes, and immediately spun at 1200g for 10 minutes at 4 °C. Plasma samples were stored at −80 °C for up to a month before assay, and only freeze-thawed once. Hormone levels were measured with specific ELISAs (ANP: no. E-EL-M0166, Elabscience Biotechnology, Houston, United States; Aldosterone: RE52301, Tecan, Hamburg, Germany) or by radioimmunoassay (Ang II: RK-002-12, Phoenix). Plasma sodium and potassium levels were determined by the Central Laboratory of the University Hospital Würzburg.

Studies of Cultured Human Adrenal H295R Cells

H295R cells (ATCC, Manassas, VA) were cultured and maintained in DMEM:F12 medium supplemented with ITS-Premix (Corning, NY), 2.5% Nu-serum, 100 U/mL penicillin and streptomycin.

Immunoblot Analyses and Gene Expression Studies

Frozen tissue samples and H295R cells were solubilized in RIPA buffer containing protease and phosphatase inhibitors. To improve the detection of the transmembrane GC-A receptor in human adrenals and adenomas, additionally, cell cytosolic and membrane proteins were separated with the Mem-PER Plus Membrane Protein-Extraction-Kit (Thermo Fisher Scientific).^16,17 Total tissue RNA was isolated with Trizol, and reverse transcription real-time PCR was performed as described.^16,17

Human Adrenal Samples

Normal adrenal cortices from patients who underwent surgery for kidney cancer, and samples from aldosterone-producing adenomas (for 1 patient together with separated samples of adjacent healthy adrenal cortical tissue) were received in collaboration with Prof. Fassnacht and Prof. Hahner at the Dept. of Endocrinology, University Hospital of Würzburg. Only central parts of tumors were used, avoiding necrotic areas and contamination with normal tissue. The clinical and pathological diagnosis was made according to established criteria.¹⁸ The patients (6 females and 2 males, 44–62 years old) had single adenomas of 1.5 to 4.3 cm in size and were registered in the European Network for the Study of Adrenal Tumor (ENS@T, www.ensat.org). Mutational analyses were not performed. The study was approved by the Ethics Committee of the University of Würzburg (approval number 88/11). Written informed consent was obtained from all patients.

Statistical Analysis

Analyses were performed using Prism software (GraphPad Prism 10). Significance was assessed using the unpaired, 2-sided t test, 1-way or 2-way ANOVA or Mann-Whitney U test as specified in individual figure legends. Statistical significance was set at P<0.05. Quantitative data are presented as mean±SEM.

Results

Generation of Genetic Mouse Models With Ablated ANP/GC-A or cGMP/PDE2A Signaling in ZG Cells

Immunoblot studies revealed strikingly greater GC-A expression in adrenal glands as compared with other mouse organs (Figure 1A). As also shown, PDE2A expression is prominent in adrenals and brain, and much lower in other tissues. Within the adrenal gland, both proteins are highly expressed in the cortex and barely detectable in the medulla (Figure 1B). To elucidate whether endogenous ANP/GC-A/cGMP signaling, through activating PDE2A, regulates aldosterone secretion, we generated mice with conditional ZG cell-restricted GC-A or PDE2A deletion using a Cre/Lox strategy. Immunoblotting and immunohistochemistry demonstrated efficient ZG deletion of either GC-A (Figure 1C) or PDE2A (Figure 1D and 1E) in double transgenic GC-A^{flox/flox;As+/Cre} mice or Pde2a^{flox/flox;As+/Cre} mice, respectively. Both KO lines were viable and fertile, and had normal Mendelian inheritance, body weight and lifespan (in comparison to respective GC-A^flox/flox or Pde2a^flox/flox littermates, as controls). The following studies first focused on the role of upstream cardiac ANP in adrenal GC-A signaling.

Figure 1.

High expression levels and efficient genetic deletions of the ANP (atrial natriuretic peptide) receptor, GC-A (guanylyl cyclase-A), or of the cGMP-stimulated PDE2A in murine zona glomerulosa (ZG) cells. A through D, Immunoblot studies. Comparison of GC-A and PDE2A expression levels between different mouse organs (A), and between dissected adrenal cortices and medullas from wildtype mice (B); (C) GC-A expression in adrenal cortices from floxed GC-A mice with/without the As-Cre^TG (n=9 per genotype); (D) PDE2A (phosphodiesterase 2A) expression in adrenal cortices from floxed Pde2 mice with/without the As-Cre^TG (n=11 per genotype). Expression levels were normalized to β-actin and calculated as x-fold from the average of respective floxed controls. E, Immunohistochemistry studies of mouse adrenal sections confirmed robust expression of PDE2A in ZG cells of control mice and its deletion in ZG PDE2A knockout (KO) mice. In B and E, Disabled homolog (Dab)2 and Laminin (Lam)β1 were used as markers for ZG cells. Data are presented as the mean±SEM. P values were calculated using the Mann-Whitney U test (C) or unpaired t test (D).

Mice with Deleted GC-A Receptor in ZG Cells Have Elevated Plasma Aldosterone Despite Increased ANP and Unaltered Renin/Ang II Levels

The main stimulus for cardiac ANP release is atrial cardiomyocyte stretch by expanded intravascular volume.⁶ Circulating ANP resets physiological volume through coordinated GC-A-mediated endothelial and renal effects.^6,8 To elucidate whether this circuit involves aldosterone regulation, firstly we compared the hormonal status of control and ZG GC-A KO mice maintained on NSD (0.2% NaCl) or an HSD (with 4% NaCl, for 2 weeks). In controls, feeding an HSD increased atrial proANP protein expression and plasma ANP levels (Figure 2A and 2B). Surprisingly, ZG GC-A KO mice on NSD had higher atrial proANP levels but unaltered plasma ANP (Figure 2A and 2B). HSD did not further raise atrial proANP expression but increased circulating ANP levels more than in control littermates, indicating enhanced ANP release from filled atrial granules (Figure 2A and 2B).

Figure 2.

Mice with deleted GC-A (guanylyl cyclase-A) in zona glomerulosa (ZG) cells have elevated plasma aldosterone levels despite increased cardiac atrial expression and release of ANP (atrial natriuretic peptide) and unaltered renal renin. Control and ZG GC-A knockout (KO) littermates received a normal-salt diet (NSD) or high-salt diet (HSD) for 14 days. A, Immunoblot: atrial pro-ANP expression levels were normalized to GAPDH and calculated as x-fold relative to controls/NSD (n=5–8 mice per group). B, ELISA: Plasma ANP levels (n=5-8 mice per group). C, qRT-PCR: Renal renin mRNA expression was normalized to GAPDH (n=6–8 mice per group). D, ELISA: Plasma aldosterone (n=8–10 mice per group). Data are presented as the mean±SEM. P values were calculated using 2-way-ANOVA.

To determine whether the renin-Ang II system was altered in ZG GC-A KO mice, renal renin expression and plasma Ang II levels were measured. Under NSD conditions, renin mRNA levels were not different between genotypes; in both, HSD reduced renin expression to a similar extent (Figure 2C). Furthermore, plasma Ang II levels were also appropriately suppressed in both genotypes following the HSD (controls NSD=304.6±33.9 pg/mL [n=5], HSD=201.6±23.2 pg/mL [n=5]; KO NSD=276.8±41 pg/mL [n=5], HSD=160.7±15.2 pg/mL [n=6]; P_diet=0.04 by unpaired t test). Under both dietary conditions, the Ang II levels of KO mice trended lower, but the difference to controls was not significant (P_{genotype and genotype×diet} >0.05 [2-way ANOVA]). In control mice, the reduction of renin and Ang II was accompanied by suppression of plasma aldosterone levels (Figure 2D). Notably, ZG GC-A KO littermates had increased aldosterone levels already under NSD conditions. After HSD, the aldosterone levels decreased (in comparison to NSD) but remained above the levels of controls (Figure 2D). Such mild hyperaldosteronism of ZG GC-A KO mice with NSD was accompanied by slight increases in plasma sodium levels (146±1.64 mmol/L; as compared with 144±1.1 mmol/L in controls [n=5]) and decreases in plasma potassium (4.94±0.08 mmol/L*; versus 5.44±0.19 mmol/L in controls (n=5; *P_genotype=0.06 by unpaired t test). Together, these results indicate that endogenous cardiac ANP, via GC-A, moderates aldosterone release. Elevation of atrial and plasma ANP levels in ZG GC-A KO mice might be a compensatory response aiming to prevent or reduce hypervolemia, a possibility which we addressed with the following studies.

ANP/GC-A Signaling in ZG Cells Blunts Increases in ABP and Intravascular Volume by Dietary Salt Load

To dissect the role of adrenal GC-A signaling in ABP regulation, ZG GC-A KO and control littermate mice were subjected to sequential tail-cuff plethysmography measurements^8,16,17 before and after 2 weeks of HSD. Under NSD conditions, ABP did not differ between the 2 genotypes (Figure 3A, left). Feeding an HSD mildly but significantly increased ABP (Figure 3A). This response was enhanced in KO mice: ABP on average increased by 4.6 mm Hg in control and by 8.2 mm Hg in KO littermates (Figure 3A, right).

Figure 3.

ANP (atrial natriuretic peptide)/GC-A (guanylyl cyclase-A) signaling in zona glomerulosa cells prevents increases in arterial blood pressure (ABP) and intravascular volume by high-salt diet (HSD). A, ABP was measured by tail cuff plethysmography during normal-salt diet (NSD) conditions and subsequently during the last 3 days of the 2-week HSD. Right: Average increase in ABP during HSD (ΔSBP=SBP_HSD–SBP_NSD) in 17 control and 16 ZG GC-A KO mice. B and C, Terminal left ventricular (LV) pressure-volume recordings in mice receiving either an NSD (B, n=6) or an HSD (C, n=8) for 2 weeks. Because these 2 experimental studies (NSD, HSD) were not performed simultaneously, the results illustrated in B and C cannot be directly compared. D and E, Morphometric analyses of cardiomyocyte cross-sectional areas (D) and interstitial collagen fractions (E) on PAS- or Sirius red-stained LV sections. P values were calculated by 2-way ANOVA (A, D, and E) or unpaired t test (B and C).

LV function was measured in mice receiving either an NSD or 2 weeks of HSD. Under both conditions, the maximal rates of LV contraction/relaxation, end-diastolic pressures, relaxation time constant (Tau, in msec), and effective arterial elastance (E_a, in mm Hg/μL) were similar in control and ZG GC-A KO mice (Table S1A and S1B). As shown, the LV ejection fractions of the KO mice were slightly reduced, although the difference from controls did not reach statistical significance. Regarding LV volume parameters, under NSD, cardiac output (in μL/min) and end-systolic as well as end-diastolic LV volumes (in μL) did not differ between the genotypes (Figure 3B). However, after HSD, the ZG GC-A KO mice had significantly greater cardiac output and end-diastolic as well as end-systolic LV volumes than their control littermates (Figure 3C). Together, these results demonstrate that LV contraction and relaxation of the KO mice were unaltered; however, under HSD, their hearts were handling more volume against mildly elevated ABP.

Mice With Deleted GC-A in ZG Cells Develop Mild LV Hypertrophy Under HSD

With NSD, control and ZG GC-A KO mice had similar LV cardiomyocyte cross-sectional areas and interstitial collagen fractions (Figure 3D and 3E). HSD mildly increased myocyte areas and interstitial collagen. The hypertrophic responses were greater in the KO mice (Figure 3D), including enhanced LV mRNA expression of the hypertrophy marker gene Nppb (encoding BNP; Figure S2). The slight profibrotic effects of HSD (Figure 3E) as well as LV collagen (Col1a1 and Col3a1) mRNA levels (Figure S2) did not differ between the genotypes.

Together, these data indicate that ZG GC-A KO mice have elevated aldosterone levels despite augmented ANP and unaltered renin/Ang II, like patients with PA.¹⁹ Under NSD conditions, compensatory pathways such as increased cardiac ANP release, through vascular and renal actions, maintain normal intravascular volume and ABP. However, under HSD, this mild hyperaldosteronism of young and otherwise healthy KO mice led to mild but significant increases in intravascular volume, ABP, and cardiomyocyte size.

ANP/GC-A Signaling in ZG Cells Moderates Growth and Aldosterone Synthase Expression

Varying dietary sodium levels markedly impact ZG morphology and activity, with dietary sodium restriction raising ZG cell growth and CYP11B2 expression through the action of Ang II, whereas sodium loading inhibits these parameters.²⁰ We hypothesized that the inhibitory effect of dietary sodium loading is partly mediated by increased cardiac ANP release and adrenal GC-A/cGMP signaling. To follow this hypothesis, we compared ZG thickness and numbers of CYP11B2-expressing ZG cells between genotypes by immunohistochemistry. In line with previous studies,²⁰ in control mice fed an HSD for 2 weeks, the area of the PDE2A/Dab2-costained ZG as well as the percent of Dab2⁺ cells expressing CYP11B2 were reduced (in comparison to NSD; Figure 4A and 4B). Strikingly, ZG GC-A KO littermate mice had mildly expanded ZG areas already under NSD, which did not shrink on HSD (Figure 4A). Moreover, they had elevated absolute and relative numbers of ZG cells expressing CYP11B2 (in comparison to controls; Figure 4B). As shown, HSD diminished the proportion of CYP11B2-expressing ZG cells, but the numbers stayed above controls, like aldosterone plasma levels (previous Figure 2D). The data confirm that varying dietary sodium levels impact adrenal glomerular morphology.²⁰ They indicate that cardiac ANP, via GC-A, moderates ZG cell growth and aldosterone synthase expression already under baseline (NSD) conditions and participates in the inhibitory effect of an HSD.

Figure 4.

ANP (atrial natriuretic peptide)/GC-A (guanylyl cyclase-A) signaling moderates zona glomerulosa (ZG) cell growth and aldosterone synthase expression. A and B, Immunohistochemistry and morphometric analyses. A, PDE2A⁺/Dab2⁺ co-stained ZG areas (in μm²) were normalized to the outer length of the respective ZG zone (marked with dotted white lines, in μm). B, The number of Dab⁺ ZG cells expressing aldosterone synthase (Dab⁺/Cyp11B2⁺/DAPI⁺) was normalized to the total number of Dab⁺/DAPI cells in the respective region (in %). In A and B, 3 random regions from each adrenal section were analyzed and averaged (n=5–6 mice per group). C, Immunoblot analyses of GC-A, cGKI (cGMP-dependent protein kinase I), and MAPK (mitogen-activated protein kinase) expression or phosphorylation in dissected adrenal cortices. The levels of phosphorylated MAPKs ERK 42/44 and p38 were normalized to the respective total levels and calculated as x-fold from the average of controls (ERK: n=9–10; p38: n=4 per genotype). Data are presented as the mean±SEM. P values were calculated using 2-way ANOVA (A and B) or unpaired t test (C).

ANP/GC-A Signaling in ZG Cells Reduces the Activity of MAPK

Activation of the MAPK (mitogen-activated protein kinases) ERK42/44 and p38 pathways participates in Ang II-stimulated growth of cultured bovine ZG cells.²¹ Conversely, ANP/GC-A/cGMP signaling, via activation of cGKI (cGMP-dependent protein kinase), counter-regulated MAPK activation and thereby growth of different types of cells in vitro and in vivo.^22–26 To elucidate whether enhanced MAPK signaling participates in the ZG expansion observed in ZG GC-A KO mice, we dissected adrenal cortices for immunoblot studies. The results show that besides PDE2A, cGKI is also highly expressed in adrenal cortices, without differences between control and ZG GC-A KO mice (Figure 4C, left). Notably, the levels of active, phosphorylated ERK42/44 and especially of phosphorylated p38 were significantly higher in ZG GC-A KO cortices than in controls (Figure 4C).

Mice With Deleted PDE2A Have Unaltered ZG Morphology and Aldosterone Levels

As already mentioned, in cultured bovine ZG cells, the inhibitory effect of the ANP/GC-A/cGMP pathway on acute ACTH/cyclic AMP-driven aldosterone secretion is mediated by PDE2A.^12,13 To elucidate whether PDE2A also participates in the chronic inhibitory effects of ANP in vivo, we generated mice with ZG-restricted deletion of PDE2A. In view of the robust expression of this PDE in ZG cells (previous Figure 1), we expected that its deletion would profoundly affect ZG morphology and function. However, under NSD conditions, the area of the Dab2⁺-stained ZG, the proportion of CYP11B2-expressing cells, the adrenal cortical levels of phosphorylated MAPK (ERK42/44 and p38) as well as plasma aldosterone levels were similar in ZG PDE2A KO and control littermates (Figure 5A through 5C). Moreover, administration of an HSD reduced the thickness of the ZG layer, the proportion of CYP11B2-expressing cells, and plasma aldosterone levels to a similar extent in both genotypes (Figure 5A and 5C). Consistently, in both genotypes, HSD provoked similar mild increases in ABP (Figure 5D). Together, these data indicate that PDE2A is dispensable for the chronic inhibitory effects of ANP/GC-A/cGMP signaling on ZG cell growth and aldosterone release.

Figure 5.

In mice with deleted PDE2A (phosphodiesterase 2A) in zona glomerulosa cells, the zona glomerulosa (ZG) thickness, the proportion of CYP11B2-expressing ZG cells, aldosterone plasma levels, and arterial blood pressure (ABP) were unaltered. A, Immunohistochemistry and morphometric analyses. Dab2⁺ areas (in μm²) were normalized to the outer length of the respective area (marked with dotted white lines; in μm); the number of Dab⁺ ZG cells expressing aldosterone synthase (Dab⁺/CYP11B2⁺/DAPI⁺) was normalized to the total number of Dab⁺/DAPI cells in the respective region (in %). Three random regions of each adrenal section were analyzed and averaged (n=5–6 mice per group). B, Immunoblot analyses of PDE2A, cGKI (cGMP-dependent protein kinase I), and MAPK (mitogen-activated protein kinase) expression or phosphorylation in dissected adrenal cortices. The levels of phosphorylated MAPKs ERK and p38 were normalized to respective total levels and calculated as x-fold from the average of controls (n=8 per genotype). C, Plasma aldosterone levels were quantified by ELISA (n=10–15 mice per group). D, ABP was measured by tail cuff plethysmography during 3 days under a normal-salt diet (NSD) and subsequently during the last 3 days of the 2-week HSD. The right panel illustrates the average increase in systolic BP during HSD (ΔSBP=SBP_HSD–SBP_NSD; n=12–13 mice per genotype). Data are presented as the mean±SEM. P values were calculated using 2-way ANOVA (A), Kruskal-Wallis-Test (C; left in D), and unpaired t test (B; right in D).

In Cultured Human Adrenal Cells, the ANP/GC-A/cGKI Pathway Inhibits Angiotensin II-Stimulated MAPK Phosphorylation

We used human adrenocortical H295R cells to study whether cGKI participates in the inhibitory effects of ANP on MAPK activity. Immunoblotting revealed GC-A, PDE2A, and cGKI expression in this commonly used cell model (Figure S3A). As an index of cGKI activity, we evaluated the phosphorylation of VASP (vasoactive stimulated phosphoprotein) at Serine 239. Treatment of H295R cells with 10 or 100 nmol/L ANP for 10 minutes increased the phosphorylation of VASP; the membrane-permeant cGKI activator 8-Br-cGMP (100 μmol/L, 10 minutes) mimicked this effect (Figure S3A). Ang II (1 nmol/L, 5 minutes) markedly enhanced the levels of phosphorylated MAPK ERK42/44 and p38 by ≈6-fold and 3-fold, respectively (Figure S3B). As shown, a short 10-minute pretreatment of the cells with 10 nmol/L ANP or 100 μmol/L 8-Br-cGMP did not change this effect. However, a longer 24-hour pretreatment significantly attenuated the acute stimulatory effects of Ang II (1 nmol/L, 5 minutes) on ERK and p38 phosphorylation (Figure S3B) as well as the slower stimulatory effects of Ang II (1 nmol/L, 24 hours) on the expression of CYP11B2 and proliferating cell nuclear antigen (a marker for cell proliferation; Figure S4).

In Human Aldosterone-Producing Adrenal Adenomas, the Expression of GC-A Is Blunted

To address the pathophysiological relevance of our experimental observations, lastly, we compared the expression of GC-A and cGMP-regulated third messenger proteins in normal human adrenal cortices and aldosterone-producing adenomas. A scheme of these pathways is shown at the top of Figure 6. Immunoblotting demonstrated that GC-A and the cGMP-regulated proteins PDE2A, PDE3A, and cGKI are expressed in healthy human adrenal cortices (Figure 6A). Notably, in the aldosterone synthase-overexpressing adenomas, the expression of GC-A was reduced by almost 80%; the expression of PDE2A trended increased, although due to the high variability these changes were not statistically significant. The expression levels of cGKI and the cyclic GMP-inhibited PDE3A were not different between normal adrenals and adenomas (Figure 6A). To improve the detection of GC-A, we additionally studied GC-A expression in membrane proteins enriched from control and tumor samples (Figure 6B) as well as from tumor samples and adjacent healthy adrenal tissue samples (as internal control) obtained from 1 additional patient (Figure 6C). The result validated GC-A downregulation in aldosterone-producing Conn adenomas (Figure 6B and C), which was associated with increased expression and phosphorylation of MAPK ERK 42/44 and p38 (Figure 6C).

Figure 6.

Diminished expression of GC-A (guanylyl cyclase-A) in human Conn adenomas. Top, left: Scheme of the ANP (atrial natriuretic peptide) signaling pathways. A and B, Expression of CYP11B2, GC-A, PDE2A (phosphodiesterase 2A), PDE3A (phosphodiesterase 3A), and cGKI (cGMP-dependent protein kinase I) in whole tissue lysates (A, n=4) and membrane fractions (B, n=6–7) from healthy human adrenal cortices and aldosterone-producing adenomas. C, CYP11B2, GC-A expression, and MAPK (mitogen-activated protein kinase) p38 as well as ERK1/2 expression/phosphorylation in 3 separate regions of an aldosterone-producing adenoma and 3 separate regions of adjacent healthy adrenal cortical tissue obtained from an additional Conn patient. Target proteins were normalized to β-actin, and values were calculated as X-fold of the average value from normal adrenal tissues. Data are presented as the mean±SEM. P values were calculated using an unpaired t test.

Discussion

This study generated novel genetic mouse models for dissecting the ANP-mediated heart-adrenal connection. Disrupting this endocrine circuit (in ZG GC-A KO mice) resulted in mildly enlarged ZG, elevated aldosterone levels already under NSD, and insufficient suppression of ZG cell growth, CYP11B2 expression, and aldosterone release by HSD. Despite this mild hyperaldosteronism, the ZG GC-A KO mice had unaltered intravascular volume and ABP under NSD. However, they showed hypervolemia and increased ABP together with mild cardiomyocyte hypertrophy after feeding an HSD. We conclude that a physiological ANP-mediated heart-to-adrenal circuit moderates ZG cell growth and aldosterone release already under baseline conditions, preventing hypervolemic hypertension in response to chronic salt load. Notably, GC-A expression is downregulated in human aldosterone-producing adenomas. Based on our experimental observations, impaired ANP/GC-A/cGMP signaling might contribute to the enhanced growth and aldosterone production of such tumor cells.

The observation that ZG GC-A KO mice only showed increased volume and ABP after feeding HSD is consistent with published data. In otherwise healthy animals, raising aldosterone levels by genetic or pharmacological methods enhanced salt sensitivity and the capacity of HSD to induce hypertension.^27–29 In contrast, increasing aldosterone in the setting of low-salt diets had little or no effect on ABP, indicating compensation by other regulating systems.^30–32 Hence, in otherwise healthy individuals or experimental animals, hyperaldosteronism promotes salt sensitivity and initiates hypertension by amplifying salt-dependent increases in sodium retention, blood volume, and cardiac output.^27–32 In fact, large population-based studies showed that long persistent PA in humans is also not always associated with increased BP.³ However, even individuals with mildly increased aldosterone show increased cardiovascular risk.^1–3 In this context, our observation that cardiac ANP moderates ZG growth and aldosterone release already under baseline NSD conditions has physiological and possibly pathophysiological relevance. The present studies were performed in young and otherwise healthy mice. Future studies will be designed to characterize the cardiovascular impact of ZG GC-A dysfunction during aging or in the presence of comorbidities.

In line with published studies,²⁰ in control mice, the administration of an HSD for 2 weeks reduced the area of the ZG as well as the proportion of CYP11B2-expressing cells. Notably, in ZG GC-A KO mice, the ZG remained expanded, and the proportion of aldosterone synthase-expressing ZG cells as well as plasma aldosterone levels were only partly suppressed. ZG is constantly self-renewing, and dietary salt impacts these dynamics, which contributes to the adjustment of aldosterone release.²⁰ Hence, in experimental animals, sodium restriction activates renin/Ang II-driven aldosterone production, at least in part, through an expansion of the ZG layer.^20,33,34 Conversely, sodium load reduced ZG cell proliferation indices and glomerular area^20,33,34 (and present studies). These responses were not associated with changes in rosette numbers, but with adaptive changes of cellular protein synthesis and proliferation within the single glomeruli.^20,34 Studies in vitro and in vivo have linked these responses to the regulatory effects of Ang II, with sodium restriction increasing and sodium load conversely decreasing Ang II levels and thereby Ang II-dependent ZG cell growth.^20,34 Notably, previous own and published studies in cultured cells and in mice with global or conditional, cell-specific GC-A deletions demonstrated that in many types of cells, ANP counteracts the growth-stimulating effects of Ang II. Such antagonistic interactions were observed in cardiomyocytes, renal epithelial and mesangium cells, fibroblasts, and vascular smooth muscle cells.^6,22–26 Our present studies for the first time reveal that endocrine-acting cardiac ANP also inhibits the growth of ZG cells, an effect which may contribute to the chronic moderation of aldosterone release and thereby to the maintenance of blood pressure/volume homeostasis.

ZG cells express 4 cGMP-regulated regulatory proteins, as possible third messengers of ANP: the PDEs 2A and 3A; and the protein kinases cGKI and cGKII^12,13,35 (and present studies). In cultured bovine ZG cells, ANP/cGMP signaling effectively counteracted acute ACTH- and Ang II-induced aldosterone secretion by activating PDE2A, which rapidly degrades cAMP and prevents PKA activation.^12,13,36 In addition, PDE2A was shown to regulate the transdifferentiation of ZG into zona fasciculata cells.³⁷ However, as shown here, mice with ablated PDE2A in ZG cells had unaltered ZG morphology and aldosterone plasma levels. Moreover, administration of an HSD appropriately suppressed CYP11B2 expression and plasma aldosterone levels and raised ABP, as in control littermates. Together, these observations implicate that PDE2A is not the main downstream messenger mediating the chronic inhibitory effects of the ANP/cGMP pathway on ZG growth and CYP11B2 expression.

In fact, in other types of cells, the inhibitory effect of ANP on hypertrophy and proliferation is mediated by the activation of cGKI, inhibiting MAPK signaling.^22–26 Conversely, genetic deletion of GC-A or cGKI led to increased ERK42/44 and p38 activities in cardiomyocytes and kidney cells.^24–26 In line with these observations, our immunoblotting studies showed increased activation of the MAPKs ERK and p38 in adrenal cortices of ZG GC-A KO, but not ZG PDE2A KO mice. Conversely, in cultured human adrenal H295R cells, ANP inhibited Ang II-induced activation of MAPK ERK and p38, and these inhibitory actions were mimicked by 8-Br-cGMP, which directly activates cGKI. This supports our hypothesis that ANP activates cGKI signaling in ZG cells and thereby moderates MAPK activity and cell growth. Notably, the inhibitory effects of ANP and 8-Br-cGMP in vitro were only observed after prolonged treatment (24 hours) and therefore might be related to genomic actions, such as the induction of mitogen-activated protein kinase phosphatase-1.^38,39

Supporting translational relevance, the present studies show that GC-A and the downstream messengers PDE2A, PDE3A and cGKI are expressed in normal human adrenal cortices and in aldosterone-producing adenomas. Notably, the expression of GC-A was markedly downregulated in the latter. Our observations in ZG GC-A KO mice suggest that blunted ANP/GC-A/cGMP signaling may contribute to the excessive growth and aldosterone production of such tumor cells. Downregulation of GC-A in Conn adenomas might be related to the hypervolemia-driven increases in plasma ANP levels reported in these patients,¹⁹ a phenomenon named ligand-dependent homologous GC-A desensitization.⁶ However, comparison of GC-A expression levels between tumor tissue and adjacent healthy adrenal cortical tissue obtained from 1 Conn patient showed that GC-A expression was selectively downregulated in the adenoma. These preliminary observations suggest that instead of circulating ANP levels, growth factors or cytokines produced within the tumor tissue, such as TGF-β,⁴⁰ locally downregulate GC-A.

Limitations of the Study

The limitation of this study includes the small number of available human adrenal tissue samples. We performed experiments with cultured human adrenal H295R cells, but these widely used cells might not entirely mimic in vivo conditions. In mice studies, we could not study circadian changes of ABP by telemetric recordings or maintain mice in metabolic cages for studies of renal electrolyte and water excretion. These procedures are not available at our university and were not covered by our animal proposal. Lastly, because the effects of ZG GC-A inhibition were evaluated in healthy young mice, further studies are necessary to examine the impact in the presence of comorbidities, such as aging or renal failure.

Perspectives

Our experimental observations demonstrate that cardiac ANP-to-adrenal cGMP signaling moderates ZG growth and aldosterone release and participates in the response to dietary salt/volume load. Conversely, impaired ANP/cGMP signaling led to expansion of the ZG and enhanced aldosterone release. Reinforcing this pathway with neprilysin inhibitors (which reduce ANP degradation⁴¹) or with novel GC-A-activating antibodies or designer NPs^42,43 may have therapeutic potential in patients with elevated aldosterone levels. Indeed, in the setting of unsuccessful therapy of PA with MR antagonists, sacubitril/valsartan (combining neprilysin inhibition with Ang II blockade) increased ANP and decreased aldosterone plasma concentrations despite enhancing renin activity.⁴¹ Moreover, subcutaneous application of a mutated (M) stabilized ANP isoform, MANP, reduced blood pressure and plasma aldosterone in patients with resistant hypertension.^43,44 Understanding the regulation of adrenal ZG cell proliferation and CYP11B2 expression by ANP/cGMP signaling may provide insights into the pathogenesis of PA and can foster the development of novel treatments.

Article Information

Acknowledgments

The authors are thankful to Erna Kleinschroth and the Histology Facility of the Institute of Anatomy and Cell Biology at the University of Würzburg for help with advice and equipment for histological analyses.

Disclosures

None.

Supplemental Material

Major Resources Tables

Table S1

Figures S1–S4

Funding Statement

This work was supported by the Deutsche Forschungsgemeinschaft (DFG grants KU 1037/13-1, KU1037/ 14-1, and CRC 1525 to M. Kuhn; NI 1301/7-1 to V.O. Nikolaev) and by the British Heart Foundation program (grant no. RG/17/6/32944 to M. Zaccolo).