Subchronic treatment with aldosterone induces depression-like behaviours and gene expression changes relevant to major depressive disorder

Abstract

The potential role of aldosterone in the pathophysiology of depression is unclear. The aim of this study was to test the hypothesis that prolonged elevation of circulating aldosterone induces depression-like behaviour accompanied by disease-relevant changes in gene expression in the hippocampus. Subchronic (2-wk) treatment with aldosterone (2 µg/100 g body weight per day) or vehicle via subcutaneous osmotic minipumps was used to induce hyperaldosteronism in male rats. All rats (n=20/treatment group) underwent a modified sucrose preference test. Half of the animals from each treatment group were exposed to the forced swim test (FST), which served both as a tool to assess depression-like behaviour and as a stress stimulus. Affymetrix microarray analysis was used to screen the entire rat genome for gene expression changes in the hippocampus. Aldosterone treatment induced an anhedonic state manifested by decreased sucrose preference. In the FST, depressogenic action of aldosterone was manifested by decreased latency to immobility and increased time spent immobile. Aldosterone treatment resulted in transcriptional changes of genes in the hippocampus involved in inflammation, glutamatergic activity, and synaptic and neuritic remodelling. Furthermore, aldosterone-regulated genes substantially overlapped with genes affected by stress in the FST. This study demonstrates the existence of a causal relationship between the hyperaldosteronism and depressive behaviour. In addition, aldosterone treatment induced changes in gene expression that may be relevant to the aetiology of major depressive disorder. Subchronic treatment with aldosterone represents a new animal model of depression, which may contribute to the development of novel targets for the treatment of depression.

Introduction

Although stress-related psychiatric disorders such as depression and anxiety disorders are under intense investigation, the exact pathophysiological mechanisms underlying these disorders remain unclear. There is substantial evidence showing a link between the dysregulation of the neuroendocrine system and the pathogenesis of depression or anxiety disorders (Abelson et al.1996; Holsboer, 2000; Jezova, 2005; Mathew, 2008). A great deal of attention has been devoted to the investigation of neuroendocrine abnormalities related to the hypothalamic–pituitary–adrenocortical (HPA) axis.

Apart from hypercortisolism, increased concentrations of circulating aldosterone have been reported in patients with depression (Emanuele et al.2005; Murck et al.2003). It can be argued that this is not an astonishing finding, given the fact that adrenocorticotropic hormone (ACTH) regulates the release of both cortisol and aldosterone (Armbruster et al.1975; Whitworth et al.1985) and therefore increased aldosterone levels could be an epiphenomenon of increased HPA axis activity without functional relevance. However, the main regulator of aldosterone release is angiotensin II, not ACTH, and the interactions between the renin–angiotensin–aldosterone system (RAAS) and the HPA axis are complex. For example, the mineralocorticoid receptor (MR), a nuclear receptor that is the physiological receptor of aldosterone, is a major component of the feedback system of the HPA axis (De Kloet et al.1998). Angiotensin II also stimulates the HPA axis by activating corticoliberin secretion at the level of the paraventricular nucleus (Jezova et al.1998).

Several observations link the RAAS to stress and stress-related disorders. Increased RAAS activity in psychosocial stress was described in healthy humans (Makatsori et al.2004) and in animal models (Grippo et al.2005). In patients with depression, several biomarkers of major depression were found to be induced by administration of the MR antagonist spironolactone, including an increase in HPA activity and a reduction of slow-wave sleep (Born et al.1991; Heuser et al.2000). On the other hand, MR antagonists exhibited antidepressant efficacy in patients with premenstrual syndrome (O'Brien et al.1979; Wang et al.1995).

MRs in most parts of the brain are occupied by glucocorticoids, which are 100–1000 times more abundant than mineralocorticoids. In peripheral tissue, a selective action of aldosterone at the MR is possible because the receptors are protected from glucocorticoids by means of the intracellular enzyme 11β-hydroxysteroid dehydrogenase type 2 (11-β-HSD2), which inactivates glucocorticoids (Stewart & Whorwood, 1994). In the brain, selectivity of MR to aldosterone is only established in a few anatomical areas, which are characterized by co-expression of MR and 11-β-HSD2. One such brain structure is the nucleus of the solitary tract (Geerling et al.2006). This region is of particular interest in stress research (Ulrich-Lai & Herman, 2009) and is an important structure mediating the antidepressant action of vagus nerve stimulation (Nemeroff et al.2006) and the effects of inflammatory cytokines on behaviour, mood and cognition (Maier & Watkins, 1998). In addition, while it is well-established that some actions of aldosterone are mediated via transcriptional regulation through MR, additional effects independent of gene regulation have been observed (Connell & Davies, 2005; Ngarmukos & Grekin, 2001).

Finally, the main function of aldosterone – the regulation of mineral metabolism at the level of the kidney – can lead to an indirect influence on CNS activity. Aldosterone is involved in the excretion of magnesium ions (Mg²⁺), which may reduce intracerebral Mg²⁺ levels and may disinhibit N-methyl-d-aspartic acid (NMDA) receptors (Murck, 2002). Interestingly, reduced Mg²⁺ content in the brain has been observed in patients with major depression compared to healthy controls by means of magnetic resonance spectroscopy (Iosifescu et al.2005). Furthermore, NMDA receptor antagonists, including ketamine, have been reported to have antidepressant activity (Berman et al.2000).

The main issue to explore is the potential causal relationship of hyperaldosteronism in the generation of stress-related disorders and, in particular, major depression. We have recently demonstrated the induction of anxiety-like behaviour in rats after subchronic aldosterone administration (Hlavacova & Jezova, 2008) and also the association of aldosterone with trait anxiety in humans (Jezova & Hlavacova, 2008). Here we investigated the potential causal role of aldosterone to induce depression-related behaviour and gene expression changes. The main hypothesis to be tested was that prolonged elevation of circulating aldosterone induces depression-like behaviours accompanied by changes in gene expression in the hippocampus.

Method

Animals

The subjects were 40 male Wistar rats (AnLab, Czech Republic) weighing 150–175 g and aged 7 wk at arrival. They were housed individually in standard cages with free access to rat chow and tap water, in a temperature-controlled room (22±2°C) under a 12-h light/dark cycle (lights on 06:00 hours). The rats were acclimated to the housing conditions for 1 wk prior to experiments and were handled daily by the experimenter. A separate group of 20 male Wistar rats weighing 250–300 g was used in the quinine taste preference test. All experimental procedures were approved by the Animal Health and Animal Welfare Division of the State Veterinary and Food Administration of the Slovak Republic and conformed to NIH Guidelines for the Care and Use of Laboratory Animals.

Experimental procedures

Treatment and osmotic minipump implantation

The animals were randomly assigned to two treatment groups (n=20 rats/group): those treated with aldosterone (d-aldosterone, A05521, Germany) and those treated with vehicle. Aldosterone and vehicle were continuously administered via osmotic minipumps for 2 wk (Alzet model 2002, 0.5 µl/h, Alza Corp., USA). Osmotic minipumps were implanted subcutaneously under combined ketamine (5% Narkamon, 1.2 ml/kg) and xylazine (2% Rometar, 0.4 ml/kg) anaesthesia to deliver aldosterone at the dose of 2 µg/100 g body weight per day. The dose of aldosterone was chosen based on our previous study demonstrating that this dose resulted in a 4-fold increase in plasma aldosterone levels (Hlavacova & Jezova, 2008). Control animals received minipumps that contained vehicle (1% ethanol solution) only. Aldosterone solubilization and implantation of osmotic minipumps have been described in detail in our previous work (Hlavacova & Jezova, 2008).

Water intake and body-weight gain measurements

Twenty-four-hour water intake was monitored daily by subtracting the weights of water containers from their respective weights on the previous day. The containers were refilled as necessary (every 3–4 d). The body weight of animals was measured once a week.

Behavioural testing and study design

Sucrose preference test

Because of significantly increased water intake in rats treated with aldosterone (Hlavacova & Jezova, 2008), we decided to use a modified version of the sucrose preference test (Rygula et al.2005). During the test, rats were given a free choice between two bottles, one with 1% sucrose solution and another with tap water for 24 h. The animals were not deprived of water and food before and during the test. To avoid possible effects of side preference in drinking behaviour, the position of the bottles was switched after 24 h. The sucrose preference test was performed in all rats for six consecutive days for days 4–10 of treatment. The consumption of sucrose solution was measured by weighing the bottles every 24 h. Due to technical problems with the measurement of water intake on the last day of the test, data from the first 5 d only were taken into consideration. The absolute sucrose intake (g) was the weighed sucrose solution uptake per rat, the relative sucrose intake (g/g body weight×100) was the absolute sucrose intake per gram of rat body weight. The percentage of sucrose solution from the total liquid ingested (sucrose preference) was considered the parameter reflecting hedonic behaviour.

Forced swim test (FST)

On days 14 and 15 of treatment, half of the animals from each treatment group were tested in the FST. The procedure was a modification of that described by Porsolt et al. (1977) . This test was simultaneously used as a stress stimulus to evaluate hormone responses. The rats were individually placed in a glass cylindrical tank (40 cm height, 14 cm diameter) filled with tap water (23±1°C) to a depth of 25 cm. Two swimming sessions were conducted. After the first 15-min session (pre-test, day 14 of treatment) rats were removed from water, dried with towels and returned to their home cages. Twenty-four hours later, rats were subjected to the 5-min test session (test, day 15 of treatment) and then immediately decapitated. Both swimming sessions were conducted during the light phase of the cycle, between 09:00 and 11:00 hours and were videotaped by camera positioned in front of the water tanks. Fresh water used for each rat and the cylinder was cleaned. Behaviour was scored over the last 4 min of the first 6 min of the pre-test session and during the entire 5 min of the test session. The percentage of time which the animal spent immobile was rated as depression-like behaviour. Rats were judged immobile when their general activity was minimized to occasional and small movements of legs or tail necessary to keep their heads above the water. The minimal duration of a bout of immobility was set at 2 s. Latency to the first bout of immobility was also scored. In addition, time which the animal spent struggling (making intense movements with forepaws in and out of the water, usually directed against the walls) and swimming (making active swimming motions, more than necessary to merely keep the head above water) was also calculated.

Quinine preference test

A separate group of rats was used to test the taste preference for bitter solutions during aldosterone treatment (n=10/treatment group). Aldosterone was administered via osmotic minipumps as described above. Quinine intake was assessed for two consecutive days from days 4–6 of treatment. During the test, rats were given a free choice between two bottles, one with 0.002% quinine (quinine hydrochloride, Sigma, Germany) solution and another with tap water for 48 h. The dose of quinine was selected according to the data of Romeas et al. (2009) . The animals were not deprived of water and food before and during the test. The position of the bottles was switched after 24 h and quinine intake was determined as described for the sucrose preference test.

Blood and organ collection

On day 15 of treatment, both control (home cages) and behaviourally tested (immediately after performing the test session of FST) rats were quickly moved to an adjacent room and euthanized by decapitation. Trunk blood was collected in cooled polyethylene tubes containing EDTA as anticoagulant and centrifuged immediately at 4°C to separate plasma, which was then stored at −20°C until analysed. The brain was quickly removed from the skull. The hippocampus was dissected, frozen in liquid nitrogen, and stored at −70°C until analysed. Adrenals were removed from the body, cleaned and weighed.

Hormone measurements

Plasma aldosterone levels and plasma renin activity (PRA) were measured by radioimmunoassay (RIA) using commercially available kits (RIA Aldosterone kit, Angiotensin I RIA kit, Immunotech, France). Plasma corticosterone levels were analysed by RIA after dichloromethane extraction as described earlier (Moncek et al.2004). Antibody for corticosterone was kindly provided by Professor C. Oliver (Marseille, France). Vasopressin concentrations in plasma were determined by specific RIA described previously (Bakos et al.2007; Jezova & Michajlovskij, 1992).

Statistical analysis of hormonal and behavioural data

Data were first checked for the normality of distributions using the Shapiro–Wilks test. Non-normally distributed data (sucrose and water intake data, hormonal data) were subjected to natural log transformations before parametric analyses were performed. Statistical analysis of data obtained from FST was performed by a t test for independent groups. The data on daily water, sucrose and quinine intake, and body weight gain were evaluated by ANOVA for repeated measures, with treatment as the between-subjects factor and time as the within-subjects factor. The data on hormone concentrations were analysed by two-way ANOVA with the factors treatment and stress. Whenever interaction reached significance, Scheffé's post-hoc test was performed. Results are expressed as original untransformed mean±s.e.m. values. The overall level of statistical significance was defined as p<0.05.

Gene expression profiling

RNA was isolated from frozen hippocampi using RNeasy mini kits (Qiagen, USA), according to the manufacturer's protocol. RNA was successfully isolated from seven vehicle-treated unstressed animals, 10 vehicle-treated stressed animals, 10 aldosterone-treated unstressed animals, and 10 aldosterone-treated stressed animals, giving a total of 37 RNA samples. The integrity of the RNA was determined using the Agilent Bioanalyzer (Agilent Technologies, USA). The RNA integrity numbers (RINs) ranged from 8.90 to 10.00 out of a scale of 10. The 37 samples from the four groups were randomized using a randomized complete block design with a block size of four and used to probe a HT_Rat-Focus 96-chip high-throughput Affymetrix gene expression array (Affymetrix, USA) using standard methods. Samples were not pooled, nor were technical replicates run. Gene expression values were normalized with the anti-log robust multi-array average (RMA) Irizarry algorithm. Genes with significantly different expression between vehicle- and aldosterone-treated animals were identified by conducting a t test on RMA-normalized intensity values using the Partek^® Genomics Suite 6.5 (Partek^® Inc., USA). A t test analysis was selected over a two-way ANOVA in order to reduce Type 2 (false-negative) errors that may obscure biologically relevant pathways. The concomitant increase in Type I (false-positive) errors was addressed by pathway analysis statistics (see below) and confirmation of representative gene expression changes by quantitative reverse transcription–polymerase chain reaction (qRT–PCR).

In order to reveal changes in expression of biologically important subsets of genes, microarray data were subjected to pathway analyses. Pathway analysis was conducted to determine whether a specified group of genes for a given biological pathway was associated with aldosterone treatment or stress (behavioural testing) exposure. The pathways were generated through use of Ingenuity^® Pathway Analysis (Ingenuity^® Systems; www.ingenuity.com). Canonical pathway analysis identified the pathways from the Ingenuity Pathways Analysis library of canonical pathways that were most significant to the dataset. Molecules from the dataset that met the p<0.05 cut-off and were associated with a canonical pathway in Ingenuity's knowledge base were considered for the analysis. The significance of the association between the dataset and the canonical pathway was measured in two ways: (1) a ratio of the number of molecules from the dataset that map to the pathway divided by the total number of molecules that map to the canonical pathway is displayed. (2) Fisher's exact test was used to calculate a p value determining the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone.

Real-time qRT–PCR

cDNA was generated by reverse-transcribing 300 ng of hippocampal RNA using the high-capacity RNA-to-cDNA kit (Applied Biosystems, USA) according to the manufacturer's protocol. qPCR was conducted in triplicate for each target gene and sample on a 7900 Fast Real-Time PCR system (Applied Biosystems) using the following cycling parameters: 1×95°C 10 min, 40×95°C 25 s, 60°C 1 min. The target genes were amplified using Taqman^® gene expression assays (Applied Biosystems, USA): prostaglandin-endoperoxide synthase 2, Ptgs2 (NM_017232.3): Rn00568225_m1 (96 bp amplicon); fatty acid amide hydrolase, Faah (NM_024132.3): Rn00577086_m1 (63 bp amplicon); interleukin 1 receptor antagonist, Il1rn (NM_022194.2): Rn00573488_m1 (76 bp amplicon); β-actin, Actb (NM_031144.2): Rn00667869_m1 (91 bp amplicon). All amplicons spanned introns. qPCR data was analysed using Sequence Detection Systems version 2.3 and RQ Manager version 1.2 (Applied Biosystems). Gene expression levels were determined using the ΔΔC_t relative quantification method. Specifically, the technical replicates of the cycle threshold (C_t) values were averaged for each sample and the average C_t for the reference gene (β-actin) was subtracted from the average C_t of the target gene to give the ΔC_t value for each sample. Coefficients of variation ranged from 0.00089–1.56% for technical replicates, and 6.05–23.86% for biological replicates. To calculate the ΔΔC_t, the average ΔC_t of the vehicle-treated animals was subtracted from the ΔC_t of the experimental sample. To calculate relative expression, the formula 2^−ΔΔC_t was used, which assumes doubling of the amplicon every amplification cycle.

Results

Basic characteristics

As revealed by ANOVA for repeated measures, aldosterone treatment did not alter body-weight gain during the experiment. Obviously, there was a significant main effect of time on body-weight gains (F_2,74=7.48, p<0.001, data not shown). No effect of treatment×time interaction was detected.

As expected, daily water intake was markedly higher in aldosterone-treated rats compared to that in vehicle-treated rats (F_1,29=4.99, p<0.05, data not shown).

Depression-like behaviours

Sucrose preference test

Aldosterone treatment induced an anhedonic state manifested by both decreased sucrose preference and sucrose consumption. The data were not normally distributed and were therefore natural log-transformed for statistical analysis. As determined by repeated-measures ANOVA, there was a significant main effect of treatment on absolute (F_1,33=7.69, p<0.01, Fig. 1) as well as relative (F_1,33=7.21, p<0.05, Fig. 1) sucrose intake showing decreased sucrose consumption in aldosterone-treated rats. In both treatment groups, there was a significant main effect of time on relative sucrose intake (F_4,132=2.61, p<0.05). Similarly, sucrose preference was significantly lower in aldosterone-treated rats compared to that in vehicle-treated rats (F_1,34=4.21, p<0.05, Fig. 1). No significant effect of time, or interaction between treatment and time was detected. Neither absolute (data not shown) nor relative water intake (Fig. 1) during the sucrose preference test was modified by aldosterone treatment.

FST

Chronic elevation in plasma aldosterone levels was associated with increased depression-like behaviour measured in the FST both in pre-test as well as in test sessions (Fig. 2). Statistical analysis by t test revealed that aldosterone-treated rats spent significantly longer time immobile (t₁₈=2.75, p<0.05) and showed significantly decreased latency to become immobile (t₁₈=−2.13, p<0.05) during the test session. Similarly, during the pre-test session, aldosterone-treated rats showed significantly (t₁₈=−2.83, p<0.05) more immobility than vehicle-treated rats (40.7±4.2% vs. 26.6±2.59%). Although there was a tendency for reduced time spent in struggling behaviour in aldosterone-treated animals, the difference did not reach statistical significance. Aldosterone treatment did not affect swimming behaviour.

Quinine preference test

Aldosterone treatment did not modify quinine intake. As the 24-h quinine intake on both test days in each rat was comparable, the values from the two test days were averaged. Statistical analysis by t test showed no differences between aldosterone- and vehicle-treated rats in absolute quinine intake (Veh: 16.3±3.65 g, Aldo: 12.78±3.91 g), relative quinine intake [Veh: 5.63±1.25 g/g body weight (BW)×100, Aldo: 4.42±1.37 g/g BW×100], relative water intake (Veh: 6.94±1.91 g/g BW×100, Aldo: 9.39±2.17 g/g BW×100) or quinine preference (Veh: 49.37±10.89% of total fluid, Aldo: 32.8±11.42% of total fluid).

Neuroendocrine response

Concentrations of all hormones in plasma were not normally distributed and were therefore natural log-transformed for statistical analysis. Two-way ANOVA of plasma aldosterone data (Fig. 3) revealed significant main effects of treatment (F_1,35=5.05, p<0.05) and stress (F_1,35=63.4, p<0.001), as well as significant treatment×stress interaction (F_1,35=18.59, p<0.001). Both aldosterone treatment and stress exposure resulted in a significant rise in plasma aldosterone concentrations. The subsequent Scheffé's post-hoc test showed a significant stress-induced increase in plasma aldosterone levels in vehicle-treated rats (p<0.001), but not in aldosterone-treated animals. In rats not exposed to stress, aldosterone treatment led to a significant rise in plasma concentrations of aldosterone (p<0.001).

Aldosterone treatment resulted in a decrease in PRA. PRA rose in response to stress induced by FST. There were significant main effects of treatment (F_1,35=21.2, p<0.001) and stress (F_1,35=42.86, p<0.001), as well as a treatment×stress interaction (F_1,35=7.12, p<0.05) on PRA (Fig. 3). Post-hoc testing showed that the stress-induced elevation in PRA was similar in aldosterone- and vehicle-treated animals. In rats not exposed to stress, aldosterone treatment resulted in a significant decrease in PRA (p<0.001).

Stress induced by FST led to a significant rise in plasma corticosterone concentrations. As determined by two-way ANOVA, there was a significant main effect of stress (F_1,35=85.92, p<0.05, Fig. 3). No effects of treatment or treatment×stress interaction were detected.

Plasma levels of vasopressin were not affected by the treatment with aldosterone. Stress induced by the FST had no significant impact on plasma levels of vasopressin. No significant treatment×stress interaction was found (Veh-control: 0.89±0.06 pg/ml; Aldo-control: 1.72±0.84 pg/ml; Veh-stress: 0.67±0.04; Aldo-stress: 0.77±0.03).

Gene expression profiling

The effect of aldosterone treatment on gene expression in the hippocampus was evaluated using Affymetrix microarrays. Probe sets (n=1411) showed significantly altered expression in aldosterone-treated animals compared to vehicle-treated animals (p<0.05 by t test), 305 with changes >20%. Ingenuity^® Pathway Analysis identified several disease-relevant processes that were significantly overrepresented in this gene set (Fig. 4, Supplementary Table 1). Many of these pathways were involved in inflammation, including complement signalling (p=3.02×10⁻³), eicosanoid signalling (p=4.43×10⁻³), dendritic cell maturation (p=1.46×10⁻²) and leukocyte extravasation signalling (p=3.232×10⁻²) (Table 2). Some changes were indicative of increased inflammation (Fig. 5), such as up-regulation of cyclooxygenase 2 (COX-2, Ptgs2, 1.2-fold, p=0.0267), up-regulation of fatty acid amide hydrolase (Faah, 1.1-fold, p=0.0098), and down-regulation of the endogenous interleukin-1 receptor antagonist, IL-1RA (II1rn, 1.1- to 1.3-fold, p=0.0353). Changes in expression levels of Ptgs2, Faah and Il1rn were confirmed by qRT–PCR (Fig. 5).

Other pathways appeared to be involved in cytoskeletal reorganization, and neuritic or synaptic remodelling (Supplementary Table 1). These include small GTPase pathways (RhoA signalling, p=1.3×10⁻³; Cdc42 signalling, p=6.993×10⁻³; Rac signalling, p=1.06×10⁻²), actin cytoskeleton signalling (p=1.29×10⁻²), axonal guidance signalling (p=9.06×10³), and ephrin receptor signalling (p=3.42×10⁻²). Many of these pathways shared aldosterone-responsive genes, such as components of the Arp2/3 complex (Actr2, Actr3, Arpc1b, Arpc2), which mediates actin nucleation and polymerization, leading to dendritic spine morphogenesis in neurons, myosin light chain regulatory subunits (Myl12b, Myl9), Cdc42 effectors (Cdc42ep3, Cdc42ep4, Cdc42ep5), and GTPase-activating proteins (Iqgap1, Iqgap3). Other genes only appeared in the neuronal pathways, such as Wnt genes (Wnt2b, Wnt5b, Wnt6, Wnt16), ephrin receptors (Ephb2, Ephb6), and semaphorins (Sema3b, Sema6c).

Glutamate receptor signalling was another pathway of interest significantly affected by aldosterone administration (p=2.61×10⁻³; Table 1 and Fig. 6). A number of glutamate receptors were altered in response to aldosterone treatment, including α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors (AMPAR1, AMPAR4), NMDA receptors (NR3A, NR3B) and metabotropic receptors (mGluR1, mGluR6). In addition, the high-affinity aspartate/glutamate transporter EAAT4, encoded by the Slc1a6 gene, was down-regulated 1.11-fold.

Hippocampal gene expression changes from animals subjected to the FST, which itself is a form of stress, were also analysed. Expression of 947 probe sets were altered by stress in the hippocampus, 129 with changes >20%. Of the probe sets altered >20% by stress, 101 (78%) – representing 83 known genes – were also altered >20% by aldosterone treatment (Table 3), demonstrating substantial overlap between stress- and aldosterone-regulated genes. In other words, of the mRNA species that changed >20% under the two conditions, 101 are in common (78% of the genes affected >20% by stress, 33% of the genes affected >20% by aldosterone). All but four of these common genes were down-regulated.

Discussion

Present results demonstrate depressogenic effects of aldosterone. Subchronic treatment with this mineralocorticoid hormone led to a depression-like behaviour, namely to behavioural despair and anhedonia. A causal involvement of aldosterone in the development of depression-like behaviour is further supported by observed changes in hippocampal gene expression and by our previous findings on the effect of aldosterone on anxiety, a common symptom of major depression.

Increased depression-like behaviour in aldosterone-treated rats has been revealed using two behavioural tests, the FST and the sucrose preference test. The behavioural data do not seem to be confounded by consequences of hypertension or muscle weakness as the dose of aldosterone used in the present experiments was not found to elevate blood pressure (Garwitz & Jones, 1982) or to produce changes in general locomotor activity in the open-field test (Hlavacova & Jezova, 2008). The FST is based on the protocol originally introduced by Porsolt et al. (1977), which is now commonly used as a tool to reveal potential antidepressant activity of new compounds. In this study, aldosterone treatment resulted in increased time spent immobile, although under the conditions of the present experiments, the control rats spent relatively little time immobile compared to our previous experiments (Mlynarik et al.2007). However, latency to first immobility, another parameter measured in the present FST, was decreased in rats treated with aldosterone, indicating that aldosterone-treated animals displayed behavioural despair much sooner than vehicle-treated ones. The measurement of latency to the first bout of immobility has recently been shown to improve the predictive validity of this behavioural despair test (Castagné et al.2009). Depressogenic effect of aldosterone was also manifested by decreased sucrose preference and sucrose intake. In rats, a reduced preference for palatable solutions indicates a decreased sensitivity to reward and may be homologous to anhedonia (Moreau, 1997). It should be noted that due to our previous findings on increased water intake following aldosterone treatment (Hlavacova & Jezova, 2008), we used a modified version of the sucrose preference test without water deprivation. Notably, aldosterone-treated rats already exhibited an anhedonic state on day 5 of treatment. It might be argued that the reduced sucrose intake could have been due to some general effects of aldosterone upon taste sensitivity. This is conceivable, but not very likely, as we failed to observe any changes in quinine preference. Thus, the present findings provide the first evidence for aldosterone-induced depression-like behaviours.

The behavioural effects of aldosterone might have been mediated via changes in other neuroendocrine functions, such as the renin–angiotensin system or HPA axis activity. As pharmacological blockade of the renin–angiotensin system was found to be associated with antidepressant effects (Gard et al.1999; Giardina & Ebert, 1989; Martin et al.1990) and aldosterone treatment caused a decrease in PRA, changes in this system do not seem to impact on the depressogenic action of aldosterone. Similarly, consistent with our previous observations (Hlavacova & Jezova, 2008), aldosterone treatment did not affect the stress-related HPA axis. Moreover, vasopressin concentrations were not influenced by treatment with aldosterone. These data support a direct effect of aldosterone itself and not an effect mediated via relevant neuroendocrine systems.

In addition to behavioural effects, aldosterone treatment induced gene expression changes in the hippocampus that may be relevant to the aetiology of major depressive disorder (MDD). Genes affected by aldosterone treatment can be grouped into three main categories: inflammation, synaptic remodelling and glutamatergic activity.

An important finding of these experiments is altered expression of genes related to inflammatory pathways. Activation of the RAAS has previously been reported to induce inflammation (Ahokas et al.2003; Duprez, 2006). For instance, COX-2 protein has been shown to be up-regulated in mouse kidney tissue following aldosterone treatment (Sontia et al.2008), consistent with our observations in this study that COX-2 message was up-regulated in the hippocampus. Inflammation has also been implicated in the pathophysiology of MDD (Dantzer et al.2008; Khairova et al.2009). For example, inflammatory genes are associated with MDD (Fertuzinhos et al.2004; Levinson, 2006), peripheral cytokines are elevated in MDD patients (Danner et al.2003; Howren et al.2009; Maes et al.1997) and treatment of patients with pro-inflammatory cytokines directly causes depressive symptoms (Kent et al.1992; Zorrilla et al.2001). Furthermore, anti-inflammatory manipulations appear to have antidepressant activity. This includes inhibitors of COX-2 (Muller & Schwarz, 2007) and polyunsaturated fatty acids (Murck et al.2004; Peet & Horrobin, 2002). Interleukin-1 (IL-1) has been implicated as one of the primary cytokines responsible for the behavioural effects of inflammation (Fertuzinhos et al.2004; Rothwell & Luheshi, 2000). Down-regulation of the gene encoding the endogenous IL-1 receptor antagonist (Il1rn) with aldosterone treatment (see Fig. 5) suggests that there is potentiated IL-1 signalling in the aldosterone model. Interestingly, genetic polymorphisms in IL-1 and IL-1 receptor antagonist have been implicated in depression (Fertuzinhos et al.2004). Taken together, these data suggest that chronic aldosterone treatment may provide a model in which it is possible to test novel anti-inflammatory therapies in depression.

Observed changes in the expression of genes involved in cytoskeletal reorganization in neurons may indicate that neuritic and synaptic remodelling is also occurring in our aldosterone model. Changes in the complement system may also be indicative of these processes, since complement has been shown to mediate synaptic pruning (Bolton & Eroglu, 2009; Pfrieger, 2009; Stevens et al.2007). There is evidence for the presence of synaptic remodelling in MDD. Chronic depression may result in reduced hippocampal volume, thought to primarily represent dendritic loss (Gurvits et al.1996; Shah et al.1998; Sheline et al.1996). As a potential consequence, MDD is associated with cognitive impairments (Austin et al.2001). Changes in cytoskeletal pathways observed in this study may also represent inflammatory pathways, such as Fcγ-mediated phagocytosis by macrophages or monocytes and leukocyte extravasation signalling.

Finally, genes related to glutamatergic pathways showed alterations in expression. This is of significant interest, since increased glutamate signalling is implicated in the pathophysiology of MDD (Kendell et al.2005). Aldosterone regulates excretion of ions, including Mg²⁺ (Sontia et al.2008). Reduced Mg²⁺ would be predicted to relieve the Mg²⁺-block on NMDA receptors in the brain, and therefore potentiate glutamate signalling. Some of the gene expression changes also predict increased glutamatergic tone. For instance, the high-affinity aspartate/glutamate transporter EAAT4, encoded by the Slc1a6 gene, was down-regulated 1.11-fold in the hippocampus. Reduced EAAT4 expression is predicted to result in reduced glutamate uptake by neurons and thus increased extracellular glutamate levels. Moreover, EAAT4 has been found to be down-regulated in MDD patients (McCullumsmith & Meador-Woodruff, 2002) and is one of the targets for the potential antidepressant drug, riluzole (Sanacora et al.2004). The AMPA receptor GluR1, encoded by the gene Gria1, has been shown to be down-regulated in Brodmann's area 21 of depression patients (Sequeira et al.2009) and was also down-regulated by the aldosterone treatment in this study. A modulator of human GRIA1 protein is a candidate therapeutic agent for MDD (Jordan et al.2005). Glutamatergic signalling is also affected by inflammation, since inflammation suppresses the ability of astrocytes to sequester glutamate and activated microglia produce the endogenous NMDA agonist, quinolinic acid (Muller et al.2006). Future studies will be needed to determine whether the changes in gene expression observed in this study translate into changes in protein expression.

Evaluation of the changes in Mg²⁺ concentrations was not in the focus of the present study. However, there is indirect evidence that aldosterone may reduce Mg²⁺ concentration in the brain because adrenalectomy was found to increase intracerebral Mg²⁺ and this effect was normalized by administration of aldosterone (Ebel et al.1971). Furthermore, Mg²⁺ depletion has been associated with inflammatory changes (Malpuech-Brugere et al.2000) and Mg²⁺ supplementation reverses inflammation induced by aldosterone (Sontia et al.2008).

One surprising outcome from the expression profiling studies is the substantial overlap between genes induced by stress and genes induced by aldosterone treatment. These results suggest that aldosterone plays an important role in the stress response and its relationship to the development of mood disorders.

In conclusion, the present study demonstrates the existence of a causal relationship between hyperaldosteronism and depression-like behaviour. Subchronic treatment with aldosterone may serve as a new model of depression. Furthermore, the hyperaldosteronism may represent an explanation for the well-recognized overlap between affective disorders and cardiovascular disease (Whang et al.2009). More importantly, the involvement of aldosterone as a mediator of changes leading to depression provides an explanation for the association of systems, which have been mainly viewed as independent. Thus, aldosterone release- or MR-related mechanisms may represent targets for further research aimed at developing new approaches leading to therapeutic benefits in patients with depression.

Note

Supplementary material accompanies this paper on the Journal's website.

Acknowledgement

The study was supported by a grant of Vega 2/0118/11.

Statement of Interest

None.

References