Cross-Talk Between Mineralocorticoid and Angiotensin II Signaling for Cardiac Remodeling

The deleterious role of aldosterone (aldo) in cardiovascular diseases has been documented by the Randomized Aldactone Evaluation Study¹ and Eplerenone Post-AMI Heart Failure Efficacy and Survival Study.² Both trials have shown a major additive benefit of the administration of mineralocorticoid receptor (MR) antagonists on top of inhibitors of angiotensin-conversion enzyme (ACE) or blockers of the angiotensin II (Ang II) type I receptor (AT1R). Pharmacological MR antagonism has also been shown to provide organ protection in several animal models (spontaneously hypertensive rats and uninephrectomized rats treated with aldo-salt, heart failure after myocardial infarction, or aortic banding). However, we know little about the mechanisms underlying the interplay involving aldo- and Ang II–induced damage in the development of myocardial remodeling and in the progression of cardiovascular diseases.

Ang II appears to be a major contributor to cardiovascular damage. There is increasing evidence that aldo, acting via MR, may mediate or even exacerbate the damaging effects of Ang II. Robert et al³ reported that administration of the AT1R antagonist losartan fully prevented both the increases in collagen I and III mRNAs and also collagen accumulation in the aldo-salt rat model, indicating cross-talk between AT1R activation and MR activation. Similar results have been reported by Sun et al.⁴ Salt loading and uninephrectomy are necessary in most models to generate both cardiac fibrosis and high circulating aldo levels, although Iglarz et al⁵ reported cardiac remodeling and fibrosis in rat infused with aldo alone. In all of the cases, cardiac fibrosis was prevented by AT1R antagonism.^3–5 This AT1R/MR cross-talk is not restricted to the heart, and MR antagonism is also beneficial in renal and vascular remodeling.⁵ Indeed, synergistic effects between aldo/MR activation and Ang II have been reported in vascular smooth muscle cells,^6,7 suggesting that this interaction is a common mechanism and may be an essential component for the effects of the renin-angiotensin-aldosterone system in the cardiovascular system.

The tissue- or cell-specific molecular and functional interactions in vivo between the components of the renin-angiotensin-aldosterone system are still poorly understood. Most studies have used experimental models, like aldo-salt treatment,^8–10 in which the increase in plasma aldo concentration is large and reaches all tissues. With excess circulating aldo, both renal and extrarenal MRs are activated, leading to numerous pathophysiological effects. These systemic effects make it difficult to determine whether the observed cardiac phenotype is because of direct cardiac MR activation or is secondary to systemic changes in, eg, sodium/potassium/magnesium homeostasis or blood pressure (BP).

Our working hypothesis is that cardiac MR/Ang II cross-talk has important molecular and functional consequences. To investigate this question, we used a transgenic mouse model with conditional cardiac-specific overexpression of the human MR. We studied cardiac function and expression of genes related to cardiac signaling during chronic Ang II administration in this model. We report additive effects between systemic Ang II and MR activation in the heart. These observations may have important consequences for the benefits of MR or combined AT1R-MR antagonist treatment in human disease.

Materials and Methods

Generation of the Human MR Conditional Model

The previously characterized tetO-hMR mouse strain¹¹ was crossed with the α-MHCtTA transactivator mouse strain (kindly provided by G.I. Fishman, New York University School of Medicine, New York) to obtain MHCtTA/tetO-hMR double transgenic (DTg) mice with conditional, cardiomyocyte-specific human MR expression. To prevent the embryonic lethality reported previously in DTg mice,¹¹ food containing doxycycline (1 g/kg) was given to the pregnant mothers until birth of the progeny to avoid human MR expression in the pups. Two-month-old DTg males were used for this study, and the littermates were used as controls (Ctls). Ang II (200 ng/kg per minute) or vehicle was infused for 2 months via osmotic minipumps (Alzet, Charles River Laboratories, Inc; Figure S1, please see the data supplement online at http://hyper.ahajournals.org). When appropriate, canrenoate (20 mg/kg per day) was administered in drinking water. The animals were euthanized at the end of treatment. The blood was collected by aorta puncture, and plasma was isolated by centrifugation for determination of plasma aldo levels. Hearts were removed, rinsed in PBS, weighed, cut into several transverse slices, frozen in liquid nitrogen, and kept at −80°C. The use of animals was in accordance with the guidelines of the European Community and approved by our institutional animal care and use committee.

BP, Urine Collection, and Echocardiographic Analysis

The BP was measured by the tail-cuff method, as published previously.¹² Twenty-four-hour urine samples were collected in individual metabolic cages (Phymep Marty Technology) for 8 days. Urine aldo concentration was determined by radioimmunoassay (Diagnostics Products). Echocardiography was performed on lightly anesthetized mice (isoflurane, Abbot, in oxygen), as described previously.¹³ Briefly, the heart was visualized in the long axis parasternal view for M-mode left ventricle (LV) dimension measurement and posterior wall pulse wave tissue Doppler measurement. An apical 4- to 5-chamber view was obtained from the subcostal view for diastolic function assessment with pulse wave spectral LV inflow and outflow and for pulse wave tissue Doppler measurement of the mitral annulus velocities.¹³

Real-Time RT-PCR and Western Blots

Detailed methods are given in the data supplement.

Ventricular Fibrosis

Ventricular samples were fixed in formalin and mounted in paraffin-embedded blocks. Sections 7-μm thick were cut, and ventricular fibrosis was measured after Sirius red staining. Image analysis was performed in a blinded fashion using IPLab software (IPLab).

Gelatin Zymography

Matrix metalloproteinase (MMP) 2 and MMP-9 activities were measured by gelatin zymography. Ventricular extracts containing 40 μg of proteins were analyzed by electrophoresis on 10% SDS-PAGE containing 1 mg/mL of gelatin (Sigma-Aldrich, United Kingdom) under nonreducing conditions, as described previously.¹⁴ Bands appeared as clear areas of lysis against a blue background and were quantified using the Multigauge software (Las3000, Fuji). Lung extract was used as a positive control.

NADPH Oxidase Activity Assay

NADPH-dependent superoxide production by 100-μg aliquots of ventricular homogenates was assayed in modified HEPES buffer (140 mmol/L of NaCl, 5 mmol/L of KCl, 0.8 mmol/L of MgCl₂, 1.8 mM of CaCl₂, 1 mmol/L of Na₂HPO₄, 25 mmol/L of HEPES, and 1% glucose [pH 7.2]) in a 96-well microplate luminometer (GloMax, Promega) in the presence of NADPH (300 μmol/L, Sigma-Aldrich) and dark-adapted lucigenin (5 μmol/L, Sigma-Aldrich). Measurements were made in duplicate. The NADPH oxidase inhibitor, apocynin (300 μmol/L, Sigma-Aldrich), was used to confirm that the assay was determining reactive oxygen species (ROS) production from NADPH oxidase.

In Situ Detection of Superoxide

ROS production in situ was detected with dihydroethidium. Cryosections were air dried at room temperature and stained with dihydroethidium (3.5 μmol/L) in PBS for 30 minutes in the dark. The slides were then rinsed with PBS and coverslipped. Analysis and image recording were performed with a Leica microscope equipped for epifluorescence. All of the images were recorded with the same time, light, and contrast parameters. Two independent observers scored the results in a blind manner for semiquantitative analysis: fluorescence was scored from 0 (no signal) to 6 (most nuclei).⁴

Oxyblot

The formation of carbonyl groups in the proteins was assessed with the Oxyblot oxidized protein detection kit, following the manufacturer’s instructions (Chemicon International). Aliquots of 10 μg of protein were used for each assay. To quantify protein oxidation, we defined an oxidation index as the ratio of the densitometric value of the Oxyblot bands to that of the corresponding bands stained with Ponceau Red (Las3000 DarkBox, Fuji Photo Film Europe, GMBH).

Statistics

Data are expressed as means±SEMs. Differences between groups were assessed with multifactor ANOVA, using JMP-6 software (SAS Institute, Inc). Values of P<0.05 were considered significant.

Results

Ang II infusion for 8 weeks induced similar increases in systolic BP in Ctl and DTg mice (Table 1). Steady-state plasma and urinary aldo levels were similar in Ang II–treated Ctl and DTg mice (Table 1). The canrenoate dose used had no effect on these variables. Both DTg and Ang II–treated control mice showed cardiac hypertrophy, as assessed from the heart:body weight ratio or heart:tibia length ratio (Table 1); however, the LV end diastolic diameter:body weight ratio estimated by echocardiography (Table 2) indicated the absence of LV dilation. The heart:body weight ratio was higher in Ang II–treated DTg mice indicating an additive effect of Ang II and MR signaling (Table 1). Ang II–treated DTg mice displayed LV hypertrophy associated with a significant left atrium (LA) enlargement (Table 2).

We studied the functional consequences of the cardiac hypertrophy associated with cardiac MR expression and Ang II challenge (Table 2). Systolic function was assessed by studying several variables, including ejection fraction and tissue Doppler imaging; there were no significant differences between the experimental groups. Ang II treatment in control mice did not change diastolic echographic variables (Table 2). In contrast, MR-overexpressing mice treated with Ang II showed an increased isovolumic relaxation time and decreased tissue Doppler diastolic wave velocity without any increase in LV filing pressure, as indicated by the unchanged maximal velocity of the LV inflow:maximal diastolic velocity of the mitral annulus ratio (Table 2); thus, they suffered from LV diastolic dysfunction. Pharmacological antagonism of MR by canrenoate prevented cardiac hypertrophy and diastolic dysfunction (Tables 1 and 2).

Then, we studied alterations of the extracellular matrix possibly associated with cardiac remodeling involved in diastolic dysfunction by analyzing interstitial collagen deposition and the mRNAs for collagen 1a (Col1a) and collagen 3a (Col3a). Histological staining with Sirius red showed that both Ang II infusion and cardiac MR overexpression caused a significant increase in cardiac fibrosis (Figure 1A and 1B). The combination of the 2 factors led to an additive increase in interstitial fibrosis (Figure 1A and 1B). The abundance of Col1a and Col3a mRNAs was greatly increased by either Ang II treatment or cardiac MR overexpression, and the combination of both led to an increase of ≈100-fold (Figure 1C and 1D). The rise in collagen gene expression was associated with a similar increase in fibronectin gene expression (Figure S2). MR antagonism with canrenoate had no effect on the Ang II–induced fibrosis or Col1a/Col3a/fibronectin expression in control mice, but it strongly reduced fibrosis (Figure 1A and 1B) and the expression of Col1a/Col3a/fibronectin (Figures 1C, 1D, and S2) in mice with combined Ang II treatment and cardiac MR overexpression. CTGF (a profibrotic factor likely involved in cardiac Ang II–induced fibrosis¹⁵) was upregulated in Ang II–treated mice and not affected by cardiac MR expression or by MR antagonism (Figure 1E). Extracellular matrix turnover, mediated through MMPs, may be involved in the observed fibrosis. Both Ang II administration and cardiac MR overexpression increased MMP-2 and MMP-9 activity assessed by gelatin zymography (Figure 1F through 1H); the effect when Ang II was administered to MR-overexpressing mice was additive, and canrenoate administration prevented the additive effect (Figure 1F through 1H). Neither ACE nor AT1R mRNA expression was modified by cardiomyocyte MR overexpression (Figure S3A and S3B), whereas AT1R expression was increased by Ang II treatment (Figure S3B).

• Download figure
• Download PowerPoint

To investigate the underlying mechanisms leading to the Ang II–worsened cardiac remodeling when cardiomyocyte MR is activated, we studied inflammation and myocardial oxidative stress in our model. Macrophage infiltration was studied by cardiac immunostaining with the F4/80 antibody directed against a macrophage epitope or by testing for CD68 gene expression: neither Ang II treatment nor cardiomyocyte MR overexpression induced macrophage infiltration (Figure S4A and S4B). The monocyte chemoattractive protein 1 inflammation marker was not increased in the heart by MR overexpression (Figure S4C). However, Ang II administration increased monocyte chemoattractive protein 1 expression, and this increase was not altered by canrenoate treatment, indicating that MR activation was not involved in this Ang II effect. Cardiac mRNA expression of several markers of inflammation, like intercellular adhesion molecule, vascular cell adhesion molecule, or tumor necrosis factor-α, was not altered by systemic Ang II or cardiomyocyte MR overexpression (Figure S4D through S4F). Importantly, the expression of inflammation markers was not enhanced by the combination of cardiac MR overexpression and Ang II treatment.

We next analyzed the expression of the NADPH oxidase subunits involved in oxidative stress. There were no differences for the expression of p47 and p67 subunits between DTg and Ctl mice (Figure S5A through S5D); p47 and p67 subunit mRNA expressions, but not protein levels, were more abundant after Ang II treatment, and this effect was not reduced by MR inhibition with canrenoate (Figure S5A through S5D). The absence of difference between Ang II–treated control and Ang II–treated MR transgenic mice (Figure S5A through S5D) indicated that there was no interaction between Ang II and MR pathways concerning these subunits. By contrast, the abundance of the gp91 (NADPH oxidase 2 [NOX2]) subunit mRNA was increased both by Ang II (20-fold increase) and by cardiac MR overexpression (12-fold increase; Figure 2A), the combination of both resulting in a 90-fold increase in gp91 mRNA (Figure 2A). Western-blot analysis showed increases, albeit of lower amplitude, in the amount of gp91 protein (Figure 2B): the combination of Ang II treatment and MR overexpression induced the largest increase in gp91 protein level, and this was partly prevented by canrenoate treatment (note that the effect of Ang II alone was not influenced by MR antagonism). The functional consequences of increased NOX2 expression were investigated through analyses of NADPH activity, superoxide production, and protein carbonylation. Maximal apocynin-inhibitable NADPH oxidase activity was similarly induced by each Ang II treatment and cardiac MR overexpression; the combination of the 2 had a greater effect than either individually (Figure 3A). Canrenoate treatment, which had only a minor effect in control mice treated with Ang II, prevented the increase in NADPH oxidase activity in Ang II–treated DTg mice, suggesting the involvement of MR. Production of superoxide, estimated using a semiquantitative approach based on dihydroethidium staining of cardiac tissue (Figure 3B), was significantly higher in Ang II–treated DTg than in Ctl mice, the effect being prevented by canrenoate administration (Figure 3B). Interestingly increased oxidative stress was associated with a 30% increase in the oxidized protein amount, as estimated by protein carbonylation (Figure 3C).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Discussion

We report evidence that interactions between systemic An II and cardiac MR-mediated pathways lead to increased fibrosis and cardiac remodeling, possibly as a consequence of a combined action on oxidative stress in the absence of inflammation. Combined increases of systemic Ang II and cardiac MR signaling were associated with LV hypertrophy and diastolic dysfunction, and this was prevented by pharmacological MR antagonism, identifying MR as a potential therapeutic target in diastolic heart failure.

The renin-angiotensin-aldosterone system is involved in the development of cardiac fibrosis,^16,17 and there is evidence implicating Ang II as a profibrotic factor.¹⁸ MR activation also favors cardiac fibrosis and remodeling in both human and animal models. A biomarker of collagen turnover, the circulating procollagen type III amino terminal peptide, is decreased by pharmacological MR antagonism in patients with severe heart failure.^19,20 Pharmacological activation of MR in the aldo-salt model results in extensive multiorgan fibrosis and remodeling,¹⁷ and AT1R activation has been proposed to contribute to the aldo-induced fibrosis.^3–5 Conversely, MR antagonism or pharmacological inhibition of aldo synthase with FAD 286 decreases organ damage associated with Ang II in the double transgenic rat overexpressing human renin and angiotensinogen.^21,22 Spironolactone treatment prevents increased cardiac interstitial fibrosis induced by Ang II in rats²³ or mice.²⁴ These various findings are consistent with there being a functional interaction between the activation of the MR and AT1R in the heart.²⁵ However, the multiple systemic changes present in the aldo-salt or aldo-infused animals make it difficult to dissect the local mechanisms responsible for this cross-talk. It is not known whether these interactions are related to direct effects of aldo via cardiac MR, to hemodynamic changes, to sodium load, or to the consequences of renal MR activation on homeostasis. Moreover, the heart is a complex organ composed of different cell types, including cardiomyocytes, fibroblasts, inflammatory cells, and vascular cells. All of them have been implicated in the cardiac effect of aldo, but the precise role of MR signaling within cardiomyocytes in the action of cardiac mineralocorticoids has not been elucidated. The combination of genetic and pharmacological approaches that we used in this study precluded confounding effects of systemic activation of MR signaling pathways, and, consequently, we were able to study the interaction between systemic Ang II and local increase in MR activation. We cannot, however, discriminate between AT1R and Ang II type 2 receptor implications. This would require further experiments using specific pharmacological antagonists.

We demonstrate that cardiomyocyte MR activation potentiates the induction of fibrosis-related genes by Ang II, with the exception of CTGF. CTGF was modulated by Ang II but not by cardiomyocyte MR overexpression, at variance from previously reported effects of the aldo/salt/uninephrectomy challenge, which may stimulate MR in other cell types than the cardiomyocyte. Increased local renin-angiotensin-aldosterone system activity has been suggested to be the mechanism by which aldo-salt induces cardiac fibrosis. Both losartan and spironolactone treatment prevent the increases in AT1R density and AT1R mRNA level, as reported by Robert et al.⁸ Increased ACE in the perivascular space and fibrotic scars, but not in cardiomyocytes, was observed by Sun et al⁴ in aldo-salt rats. The expression of ACE and AT1R in DTg mice was the same as in controls, suggesting that downstream events secondary to chronic activation of cardiac MR and Ang II receptors are involved in this cross-talk. Jaffe and Mendelsohn⁶ showed that Ang II may activate the MR in vascular smooth muscle cells via AT1R, but the pathways are unknown. Our results suggest that this mechanism may also operate in cardiomyocytes.

It has been proposed that the mechanisms by which mineralocorticoids induce cardiac fibrosis involve the induction of oxidative stress,²⁶ whereas the profibrotic effects of Ang II involve the redox-sensitive signaling pathway.^24,27 MR blockers decrease Ang II–induced inflammation, fibrosis, and oxidative stress.²⁴ MR blockade, independent of systolic BP reduction, also improves oxidative stress–induced structural and functional changes associated with an increase renin production (ren2 transgenic rat model).²⁸ Therefore, this pathway may also be involved in cardiac MR/Ang II cross-talk. We report a cumulative effect of Ang II and MR overexpression on the expression of the main NADPH oxidase subunit, gp91/NOX2, without increasing the expression of the p67 and p47 subunits. The gp91/NOX2 includes flavin-adenine-dinucleatide– and NADPH-binding domains in its transmembrane regions and is essential for the activity of the NOX2-containing NADPH oxidase. NOX2 is responsible for the Ang II–induced cardiac hypertrophy and aldo-induced cardiac fibrosis, as elegantly demonstrated using NOX2 knockout mice.²⁴ NOX2 is also expressed in cardiomyocytes,²⁹ so the increase in NOX2 subunit expression that we report may be the link between cardiac MR activation and fibrosis. It has been suggested that the aldo-induced oxidative stress originates from inflammatory cells.³⁰ One of the original findings of our work is that the cardiomyocyte MR can transduce a signal that led to increased oxidative stress, without detectable inflammatory processes. This finding in vivo confirms and extends previous data showing that MR activation stimulated the generation of ROS in isolated adult rat ventricular cardiomyocytes.³¹

Another original finding of our in vivo study is that the molecular alterations induced by the combination of Ang II and cardiomyocyte MR overexpression have important functional consequences mostly affecting LV diastolic function and LA remodeling with LA enlargement. This is reminiscent of clinical data indicating that pharmacological MR antagonism has beneficial effects on diastolic heart failure. The Resource Utilization Among Congestive Heart Failure clinical trial showed recently that more than half of heart failure patients have diastolic heart failure (DHF).³² The major modification in patients with DHF is diastolic dysfunction, including alterations of relaxation and filling, whereas systolic LV function is normal. It has been proposed that the development of cardiac fibrosis may account for the development of diastolic dysfunction in DHF patients, with the functional alteration typically described in infiltrative cardiomyopathies, such as sarcoïdosis or Fabry disease.³³ Our study showed that MR might be important in LA remodeling, because LA enlargement was observed in MR mice treated with Ang II and was prevented by MR blockade. Our study, however, cannot differentiate between the functional consequences of the greater LV hypertrophy induced with Ang II and cardiac MR overexpression and a direct effect related to the molecular consequences of the combination of Ang II treatment and cardiac MR overexpression. Nevertheless, blockade of MR, Ang II, or of both could be of clinical benefit in a situation in which LV hypertrophy is associated with diastolic dysfunction. The role of MR activation in DHF has been assessed in 2 clinical studies, but both included only small samples. Mottram et al³⁴ demonstrated that MR antagonism by spironolactone, in the absence of hemodynamic effects, improved LV function in hypertensive patients with DHF. Similar findings were reported by Orea-Tejeda et al³⁵ in DHF patients: they report favorable effects of spironolactone treatment on top of standard therapy, including ACE inhibitors and AT1R antagonists. Mineralocorticoid signaling has been implicated recently in the transition to heart failure in elderly dogs with hypertensive heart disease associated with diastolic dysfunction.³⁶ Most interestingly, increased expression of MR in cardiac tissue was reported in a hypertensive DHF experimental model (Dahl-sensitive rats fed an 8% NaCl diet); this increased cardiac MR level may account for the beneficial effect of eplerenone on diastolic function reported.³⁷ Our findings clearly support the involvement of MR in diastolic dysfunction.

Clinical Perspectives

Our study provides new insights into the involvement of MR in early cardiac alterations and particularly into the functional consequences of LV fibrosis and hypertrophy with diastolic dysfunction and LA remodeling. Therapeutic MR blockade may decrease both LV diastolic dysfunction and LA remodeling, possibly preventing the occurrence of severe complications.

Footnotes