Corticosteroids Mediate Heart Failure-Induced Depression through Reduced σ1-Receptor Expression

Cardiovascular diseases are risk factors for depression in humans. We recently proposed that σ1 receptor (σ1R) stimulation rescued cardiac hypertrophy and heart failure induced by transverse aortic constriction (TAC) in mice. Importantly, σ1R stimulation reportedly ameliorates depression-like behaviors in rodents. Thus, we hypothesized that impaired σ1R activity in brain triggers depression-like behaviors in animals with cardiovascular disease. Indeed, here we found that cardiac hypertrophy and heart failure induced by TAC were associated with depression-like behaviors concomitant with downregulation of σ1R expression in brain 6 weeks after surgery. σ1R levels significantly decreased in astrocytes in both the hippocampal CA1 region and dentate gyrus. Oral administration of the specific σ1R agonist SA4503 (0.3–1.0mg/kg) significantly improved TAC-induced depression-like behaviors concomitant with rescued astrocytic σ1R expression in CA1 and the dentate gyrus. Plasma corticosterone levels significantly increased 6 weeks after TAC, and chronic treatment of mice with corticosterone for 3 weeks elicited depression-like behaviors concomitant with reduced astrocytic σ1R expression in hippocampus. Furthermore, the glucocorticoid receptor antagonist mifepristone antagonized depressive-like behaviors and ameliorated decreased hippocampal σ1R expression in TAC mice. We conclude that elevated corticosterone levels trigger hippocampal σ1R downregulation and that σ1R stimulation with SA4503 is an attractive therapy to improve not only cardiac dysfunction but depression-like behaviors associated with heart failure.


Introduction
The prevalence of depression in heart failure (HF) patients has drawn attention because it is associated with poorer clinical outcomes [1,2]. Indeed, depression following myocardial infarction (MI) is associated with higher morbidity and mortality [3]. Current American College of Cardiology/American Heart Association guidelines for coronary artery bypass graft (CABG) surgery, acute MI, and chronic angina all recommend evaluation of depression and subsequent therapy [4]. Notably, administration of selective serotonin reuptake inhibitors (SSRIs) to MI patients reduces post-MI morbidity and mortality [3]. However, it remains unclear whether MI causes depression, because the relationship of the heart to emotional responses is largely unknown.
Since selective serotonin reuptake inhibitors (SSRIs) such as fluvoxamine and sertraline are potent σ 1 R ligands, we recently focused on cardioprotective mechanisms of SSRI through the σ 1 R in heart failure models using transverse aortic constriction (TAC) mice and ovariectomised rats [5][6][7]. We found that fluvoxamine, which exhibits a high σ 1 R affinity (Ki = 36 nM), but not paroxetine, which binds with lower affinity (Ki = 1893 nM), markedly rescued heart failure and cardiac dysfunction in both TAC models of mice and rats [5,7]. In addition, we previously confirmed that the selective σ 1 R agonist SA4503 (Ki = 4.4 nM) antagonizes cardiac dysfunction associated with heart failure in mice [8]. Interestingly, σ 1 R is highly expressed in neonatal rat cardiomyocytes and in brain [8,9]. σ 1 R protein levels in left ventricle markedly decreased accompanied by progression of left ventricular hypertrophy in TAC mice and σ 1 R protein expression was significantly and positively correlated with impaired LV fractional shortening [7]. σ 1 R protein is predominantly localized to the mitochondria-associated ER membrane (MAM) where it functions as molecular chaperone of the IP 3 receptor (IP 3 R), thereby regulating Ca 2+ transport into mitochondria from the ER [10]. In cultured cardiomyocytes, σ 1 R stimulation inhibits angiotensin II-induced hypertrophy and cardiomyocyte injury by promoting mitochondrial Ca 2+ transport and ATP production [8]. σ 1 R knockdown by siRNA totally eliminates σ 1 R-mediated cardioprotection in cultured cardiomyocytes, suggesting that SSRI or SA4503 directly acts on σ 1 R expressed in cardiomyocytes.
Recently, Ito et al. reported that pressure overload-and high salt diet-induced heart failure elicits depression-like behaviors in mouse with concomitant reduction of σ 1 R levels in brain, and that sympathetic nervous system suppression by administration of a σ 1 R agonist in the CNS antagonized both σ 1 R downregulation and impaired LV fractional shortening [11]. However, how mechanistically σ 1 R is downregulated and in what cells remains unclear.
Here, we report that TAC surgery induces depression-like behaviors evident 6 weeks later and that σ 1 R stimulation with the high affinity σ 1 R agonist SA4503 ameliorates those behaviors. We also show that σ 1 R is preferentially downregulated following TAC in hippocampal astrocytes. Depressive-like behaviors and decreased σ 1 R levels in TAC mice were ameliorated by treatment with the glucocorticoid receptor inhibitor mifepristone. Finally, we demonstrated that SA4503 treatment ameliorates TAC-induced depression-like behaviors by rescuing σ 1 R proein expression in brain and heart. We conclude that σ 1 R agonists such as fluvoxamine and SA4503 are potential therapeutics to treat depression following cardiovascular disease.

Animals and Operations
All procedures for handling animals complied with the Guide for Care and Use of Laboratory Animals and were approved by the Animal Experimentation Committee of Tohoku University Graduate School of Pharmaceutical Sciences. Adult male ICR mice weighing 35 to 40 g were obtained from Nippon SLC (Hamamatsu, Japan). Ten-week-old males were acclimated to the local environment for 1 week, which included housing in polypropylene cages at 23 ± 1°C in a humidity-controlled environment maintained on a 12-h light/dark schedule (lights on 8:00 AM-8:00 PM). Mice were provided food and water ad libitum. Transverse aortic constriction (TAC) was performed as described [7] on male ICR mice under anesthesia using a mixture of ketamine (100 mg/kg, i.p.) (Daiichi Sankyo Pharmaceutical Co. Ltd, Tokyo, Japan) and xylazine (5 mg/kg, i.p.) (Sigma, St. Louis, MO). Adequate depth of anesthesia was confirmed by a negative toe-pinch reflex. If anesthesia was not sufficient, then a top-up dose of 20% of the initial dose was given. . SA4503 and NE-100 were dissolved in a 0.9% saline solution. Paroxetine was dissolved in 0.5% carboxymethylcellulose. Vehicle, SA4503 and NE-100 (1.0 mg/kg) were administered orally (once daily) over the last 4 weeks of TAC using a metal gastric zonde for mice in a volume of 1ml/100g body weight, starting from the onset of aortic banding. To define the effect of corticosterone (CORT) on σ 1 R expression, ICR mice were randomly separated into three experimental groups: 1) vehicle (n = 8), 2) CORT for 1 week (n = 11), and 3) CORT for 3 weeks (n = 8). CORT (20 mg/kg, s.c.) was was dissolved in 0.5% carboxymethyl cellulose solution and injected for 1 or 3 weeks (once daily) in a volume of 1ml/100g of body weight as described [13]. To investigate mifepristone effects, ICR mice were separated into four groups: 1) sham plus vehicle (n = 6), 2) sham plus mifepristone (50 mg/kg s.c.) (n = 7), 3) TAC plus vehicle (n = 4), and 4) TAC plus mifepristone (50 mg/kg, s.c.) (n = 4). Mifepristone was dissolved in PEG400 vehicle (Nacalai Tesque, Inc., Kyoto, Japan). Vehicle and mifepristone were injected subcutaneously daily for 4 weeks starting 2 weeks after TAC surgery.

Tail Suspension Test
The tail suspension test is widely used in mice to test anti-depressant activity. The test is based on the fact that animals subjected to the short-term, inescapable stress of being suspended by the tail will develop an immobile posture. The depressed mice show increased immobility time. Immobility duration following tail suspension was scored according to the method of Steru et al. [14]. In brief, acoustically and visually isolated mice were suspended 50 cm above the floor by adhesive tape placed approximately 1 cm from the tip of the tail. Immobility time was recorded during a 5-min period. Mice were considered immobile only when they hung passively and remained completely motionless.

Forced Swim Test
The forced swim test is also widely used to assess antidepressant-like activity. Mice were placed individually in glass cylinders (height: 25 cm, diameter: 16 cm) filled 20 cm of water at 24°C, and the duration of immobility time in a 5-min test was scored as described by Porsolt et al. [15]. Mice were judged ''immobile" when they ceased struggling and remained floating motionless in the water, making only movements necessary to keep their head above water.

Sucrose Preference Test
Mice were randomly placed into cages with two bottles containing drinking water and 1% sucrose solution. Bottle's weights were measured once in each day and location was switched in the next day to prevent bias from preference for location. Tests were conducted for five concecutive days, and sucrose preference (%) was calculated as ratio of 1% sucrose consumption to total consumption.

Echocardiography and Measurement of Cardiac Hypertrophy
Noninvasive echocardiographic measurements were performed as described [7,8]. Briefly, procedures were performed in mice anesthetized with pentobarbital-NaCl/EtOH solution (0.05 g/ kg, i.p.) using ultrasonic diagnostic equipment (SSD-6500; Aloka, Tokyo, Japan) equipped with a 10-MHz linear array transducer (UST-5545; Aloka). The heart was imaged in a two-dimensional parasternal short-axis view, and an M-mode echocardiogram of the midventricle was recorded at the level of the papillary muscles. Diastolic and systolic LV wall thickness, LV enddiastolic diameter (LVEDD) and LV end-systolic diameter (LVESD) were measured, all from leading edge to leading edge according to American Society of Echocardiography guidelines. The percentage of LV fraction shortening (LV%FS) was calculated as ([LVEDD-LVESD]/ LVEDD) x 100. After animals were sacrificed by cervical spine fracture dislocation, the thoracic cavity was opened, and hearts were immediately harvested and weighed. Blood pressure was measured using the tail-cuff method (BP-98A-L, Softron, Tokyo, Japan), according to the manufacturer's protocol.

Western Blot Analysis and Measurement of ATP Content
Frozen samples were homogenized as described [8]. Western blot analysis was performed as previously described [7,16]. Blots were subjected to the ECL immunoblotting detection system (Amersham Biosciences, U.K.), and emitted light was captured by a Luminescent Image Analyzer (LAS-4000 mini, Fuji Film, Tokyo, Japan) attached to a CCD camera. Densitometric quantification of blot signals was performed with Image Gauge Ver3.0 (Science Lab, Fuji Film, Tokyo, Japan). ATP measurement was performed using an ATP assay kit (Toyo Ink, Tokyo, Japan), according to the manufacturer's protocol as described [17,18]. Briefly, frozen samples were homogenized in homogenate buffer (0.25M sucrose, 10mM HEPES-NaOH, pH 7.4), and lysates were cleared by centrifugation at 1000xg for 10 min at 4°C. Supernatants were collected, and supernatant proteins were solubilized in extraction buffer. After a 30 min incubation, luciferin buffer was added to each sample and oxyluciferin was detected using a luminometer (Gene Light 55, Microtec, Funabashi, Japan).

Measurement of Corticosterone Levels in Plasma and Brain Tissue
Corticosterone measurement was performed using a corticosterone EIA kit (Enzo Life Sciences, Farmingdale, NY), according to the manufacturer's protocol. Plasma samples were collected as previously reported [19]. Briefly, the mice were decapiyated and then blood was collected into tubes that contained EDTA. After centrifugation at 4°C, 1500×g for 10 min, plasma was aliquoted and stored at -80°C until CORT analysis.

Statistical Analysis
Values are represented as means ± standard error of the mean (S.E.M.). Comparison between two experimental groups was made using the unpaired Student's t test. Other results were evaluated for differences using two-or one-way analysis of variance (ANOVA), followed by multiple comparisons using Dunnett's test. P<0.05 was considered statistically significant.

Time Course Analysis of Depression-Like Behaviors and σ1R Expression
To define σ 1 R function in TAC-induced depression-like behaviors, we first evaluated temporal changes in both depression-like behaviors and brain σ 1 R expression in mice after TAC or sham surgery. In TAC animals, behaviors indicative of depression were observed following analysis of forced swimming and tail suspension tasks 6 weeks after TAC (P<0.01 vs. sham in forced swimming task and P<0.05 vs. sham in tail suspension task) (Fig 1A and 1B). Importantly, σ 1 R expression in the hippocampal CA1 region and dentate gyrus (DG) but not in medial prefrontal cortex (mPFC) region markedly decreased both at both 4-and 6-week times points after TAC (P<0.05 vs. sham in CA1 at 4 and 6 weeks after TACs, and P<0.01 vs. sham in DG at both 4 and 6 weeks) (Fig 1C and 1D). Temporal changes in σ 1 R levels were comparable to those previously reported in the heart [7].
Since σ 1 R stimulation with the novel σ 1 R agonist SA4503 reportedly improves depressionlike behaviors in olfactory bulbectomized rats [22], we next tested potential anti-depression effects of SA4503 in TAC mice. As expected, SA4503 treatments starting 2 weeks after TAC and continuing for 4 weeks dose-dependently antagonized increased immobility times, which are indicative of depression, seen in the forced swimming task (P<0.01 vs. TAC-vehicle for SA 0.3 and SA 1.0) ( Fig 1E). In tail suspension task, SA4503 treatment (1.0 mg/kg) reduced immobility times to identical levels to sham mice ( Fig 1F). However, mice treated with both SA4503 and NE100 showed similar increase in immobility times with sham mice (P<0.01 vs. sham for SA 1.0+NE in forced swimming task, P<0.05 vs. sham for SA 1.0+NE in tail suspension task) (Fig 1E and 1F). Treatment with paroxetine (0.4 mg/kg), a conventional SSRI which has a relatively low affinity of σ 1 R [23], partially but significantly reduced immobility times of TAC mice in forced swimming task (63.4 ± 5.9 sec) (P<0.05 vs. TAC-vehicle), but not in tail suspension task (102.3 ± 10.8 sec)(S1 Fig). In sucrose preference test, sucrose preference was significantly decreased in TAC mice (P<0.01 vs. sham) (Fig 1G). SA4503 treatment recovered sucrose preference in TAC mice (P<0.05 vs. TAC veh) and the effect was blocked by coadministration of NE100 ( Fig 1G). Overall, these findings indicate that SA4503 treatment ameliorates TAC-induced depression-like behaviors in TAC mice.

Effect of SA4503 Treatment on Hippocampal σ1R Expression
Since σ 1 R stimulation by SA4503 ameliorated TAC-induced depression-like behaviors, we next asked whether treatment of mice with SA4503 restored brain σ 1 R expression using the same experimental design employed for behavioral tests. SA4503 administration at 1 mg/kg significantly restored σ 1 R levels in both CA1 (P<0.05 vs. TAC-vehicle for SA 1.0) and DG regions (P<0.01 vs. TAC-vehicle for SA 1.0) (Fig 2A and 2B). Co-administration of NE-100 failed to abolish SA4503-induced increases in σ 1 R expression (Fig 2A and 2B). We also analyzed expression levels of another ER chaperone protein, 78 kDa glucose-regulated protein (GRP78, Bip), which makes a complex with σ 1 R [10]. As previously reported by Tagashira et al. [8], SA4503 administration at 0.3 and 1 mg/kg significantly rescued σ 1 R downregulation seen in heart following TAC (P<0.01 vs. TAC-vehicle for SA 0.3 and SA 1.0) (Fig 2C and 2D). In that study we also found that SA4503 treatment immediately after surgery increased heart ATP production concomitant with improved heart function [8]. In the present study, we found that SA4503-mediated restoration of ATP production also occurred in hippocampus (P<0.05 vs. TAC-vehicle for SA 1.0) (Fig 2E), and coadministration of NE-100 abolished this effect (P<0.01 vs. SA 1.0) (Fig 2E). We also confirmed that SA4503 treatment restored ATP production in the heart (P<0.05 and P<0.01 vs. TACvehicle for SA 0.3 and 1.0 mg/kg, respectively) ( Fig 2F). Taken together, SA4503-induced restoration of σ 1 R and ATP levels positively correlates with its anti-depressive action.

Immunohistochemical Changes in Hippocampal σ1R Expression
We next defined hippocampal regions in which σ 1 R is downregulated using immunohistochemistry. In CA1, σ 1 R was predominantly expressed in astrocytes (GFAP-positive cells) in the molecular layer and faintly in pyramidal neurons (see DAPI-positive cells; Fig 3A). Accordingly, σ 1 R immunoreactivity in both astrocytes and neurons in CA1 significantly decreased by 6 weeks after TAC (Fig 3A). On the other hand, when we assessed the number of living neurosn by DAPI staining, TAC had no effect on the total number of neurons in CA1 region at 6 weeks after TAC (71.17±4.64 cells (sham) vs. 70.00±3.11 cells (TAC 6 weeks), respectively (0.35 μm×0.35 μm region)). We also confirmed that the number of GFAP-positive cells in CA1 was comparable in both TAC and sham-operated mice (16.5±1.71 cells (sham) vs. 12.25±0.85 cells (TAC 6 weeks), respectively (in a 0.35 μm×0.35 μm region)). Likewise, in the DG, σ 1 R was predominantly expressed in astrocytes in both molecular layer and hilus regions of the hippocampus. σ 1 R levels markedly decreased by 6 weeks after TAC (Fig 3B). Finally, the total number of GFAP-positive cells in DG did not change in TAC mice (23.5±2.02 cells (sham) vs. 21.0 ±2.27 cells (TAC 6 weeks), respectively (in a 0.35 μm×0.35 μm region)).
Since σ 1 R expression decreased primarily in hippocampal astrocytes (Fig 3A and 3B), we confirmed the effect of SA4503 treatment on σ 1 R expression using immunohistochemistry. SA4503 treatment of mice subjected to TAC rescued σ 1 R expression primarily in CA1 astrocytes, not in pyramidal neurons (Fig 3A). Co-administration with NE-100 did not eliminate SA4503-mediated increases in astrocytic σ 1 R expression (Fig 3A). In DG, SA4503 treatment also restored σ 1 R expression primarily in astrocytes (Fig 3B), and NE-100 co-administration did not eliminate these effects (Fig 3B). Taken together, SA4503-induced restoration of hippocampal σ 1 R levels preferentially occurs in astrocytes.

Effect of SA4503 Treatment on Corticosterone Levels in Plasma and Hippocampal Tissue
We next addressed potential causes of hippocampal σ 1 R donwnregulation following TAC. Increases in plasma corticosterone (CORT) levels reportedly occur in depressed patients and are a factor in stress-induced depression [24][25][26]. Thus, we investigated CORT levels following TAC in plasma and hippocampal tissue by ELISA. As expected, plasma CORT levels increased in TAC mice at 4-and 6-week time points after TAC surgery (P<0.01 vs. sham for 4 and 6 weeks TAC, respectively) ( Fig 4A). In hippocampal tissues, however, CORT levels were not significantly changed but tended to increase in TAC mice (1.11 ± 0.11 ng/mg tissue (sham) vs. 1.35 ± 0.16 ng/mg tissue (TAC 6 weeks) (P = 0.51) respectively.) We next investigated σ 1 R expression in a CORT-induced depressive mouse model (Fig 4B-4G). CORT administration over a 3-week period increased immobility time in forced swimming (P<0.05 vs. vehicle) ( Fig  4B). CORT treatment slightly increased immobility time in tail suspension task, however, there was no stastically significance ( Fig 4C). Most importantly, CORT administration decreased σ 1 R expression at 1-and 3-week time points in hippocampus, including CA1 and DG (P<0.05 vs. control for 1 week CORT in CA1 region, and P<0.01 vs. control for 3 week CORT in CA1 region and for both 1 and 3 weeks CORT in DG region) (Fig 4D and 4E).
We also asked whether hippocampal σ 1 Rs are downregulated after CORT administration using immunohistochemistry. In CA1, endogenous σ 1 R expression markedly decreased at 1 and 3 week time points after CORT treatment in both astrocytes and neurons (Fig 4F). Likewise, in the DG, expression of endogenous σ 1 Rs in astrocytes also markedly decreased at 1 and 3 week time points following CORT treatment (Fig 4G). These findings suggest that elevated plasma CORT levels underlie TAC-induced σ 1 R downregulation, especially in hippocampal astrocytes. Furthermore, we analyzed σ 1 R expression levels in heart left ventricles. CORT Transverse Aortic Constriction Leads to Depression-Like Behaviors treatment for both 1 and 3 weeks failed to alter σ 1 R expression in heart (veh: 100 ± 6.6%, CORT 1 week: 88.1 ± 7.1%, and CORT 3 week: 107.9 ± 9.3%) (n = 5-6). These results indicate that increased plasma CORT in TAC mice triggers σ 1 R reduction in the hippocampus but not in the heart.

Effect of Mifepristone on Depression-Like Behaviors and σ 1 R Expression in TAC Mice
Next, we asked whether pharmacological inhibition of glucocorticoid receptors would eliminate TAC-induced depressive-like behaviors and/or decrease hippocampal σ 1 R expression. Treatment on TAC mice with mifepristone slightly antagonized increased immobility times seen in forced swimming (P = 0.38) (Fig 5A), however completely abolished TAC-induced increase in immobility time in tail suspension task (P<0.05 vs. TAC-vehicle for TAC-mifepristone) (Fig 5B). Moreover, mifepristone treatment of TAC mice significantly rescued reduced σ 1 R expression seen in hippocampal CA1 and DG (P<0.01 and P<0.05 vs. TAC-vehicle for TAC-mifepristone in CA1 and DG, respectively), but had no effect on σ 1 R expression in sham Transverse Aortic Constriction Leads to Depression-Like Behaviors mice (Fig 5C and 5D). These results suggest that elevated CORT levels partially function in depression-like behaviors and decreased hippocampal σ 1 R expression in TAC mice.

SA4503 Treatment Ameliorates Myocardial Dysfunction in TAC Mice
We previously documented that SA4503 administration immediately after TAC surgery rescues cardiac dysfunction [8]. Here, we confirmed those findings when SA4503 as administered starting at 2 weeks after surgery. Like our previous findings [8], TAC mice showed decreases in LV fractional shortening (FS) and increases in left ventricular end-systolic diameter (LVESD) and left ventricular end-diastolic diameter (LVEDD) (P<0.01 vs. sham) compared with shamoperated mice (Fig 6A-6D). SA4503 treatment dose-dependently rescued decreased FS (P<0.05 vs. TAC-vehicle for SA 0.3 and SA 1.0) (Fig 6B) and increased LVESD (P<0.05 vs. TAC-vehicle for SA 0.1, SA 0.3 and SA 1.0) (Fig 6C). Co-administration of NE-100 abolished SA4503-mediated amelioration of cardiac dysfunction, as indicated by decreased FS and increased LVESD (P<0.01 vs. SA 1.0) (Fig 6B and 6C). Also, TAC mice treated with vehicle, both SA4503 and NE100 showed significant increase in LVEDD (P<0.01 and P<0.05 vs. sham for TAC-vehicle and SA 1.0+NE) (Fig 6D). SA4503 treatment significantly and dose-dependently rescued the HW-to-BW ratio (P<0.01 vs. TAC-vehicle for SA 1.0) (Fig 6E). Co-administration of NE-100 abolished SA4503 inhibition of TAC-induced hypertrophy, as indicated by HW-to-BW (P<0.01 vs. TAC-vehicle for SA 1.0) (Fig 6E). We conclude that the anti-pathological hypertrophic effect of SA4503 on TAC mice is mediated by σ 1 R stimulation in vivo. Transverse Aortic Constriction Leads to Depression-Like Behaviors SA4503 Treatment Affects Sympathetic Nervous Activity as Indicated by Peripheral Blood Pressure Ito et al. previously reported that suppression of sympathetic nerve activity by a different σ 1 R agonist, PRE084, underlies improvement in impaired LV fractional shortening in a pressure overload-and high-salt diet-induced HF mouse model [11]. Therefore, we asked whether TAC surgery and SA4503 treatment altered sympathetic nervous activity. At a 6-week time period, heart rate (HR) was comparable with or without TAC surgery and with or without chronic SA4503 treatment over the last 4-week period (Fig 7A and 7B). Although mean blood pressure (MBP), diastolic blood pressure (DBP) and systolic blood pressure (SBP) significantly decreased in some points among 6 weeks following TAC surgery (MBP; P<0.01 and P<0.05 vs. sham at 2-and 6-weeks after TAC, DBP; P<0.01 vs. sham at 2-weeks after TAC, SBP; P<0.01 and P<0.05 vs. sham at 2-and 6-weeks after TAC) (Fig 7C, 7E and 7G), chronic treatment with SA4503 over the last 4 weeks did not alter MBP, DBP andSBP (Fig 7D, 7F and 7H). These data suggested that SA4503 at doses used in this study has no effect on sympathetic activity.

Discussion
Here, we showed that: 1) mice subjected to a TAC-induced hypertrophy model exhibit depression-like behaviors after a prolonged period (6 weeks); 2) these behaviors are closely associated with reduced σ 1 R expression in hippocampal astrocytes; 3) treatment of mice with the selective σ 1 R agonist SA4503 ameliorates TAC-induced depression-like behaviors with concomitant improvement of heart function; and 4) increased plasma CORT levels are among causative factors underlying σ 1 R downregulation and depression-like behaviors following TAC. Taken together, we provide novel evidence for crosstalk between heart failure and depression through the hypothalamic pituitary-adrenal (HPA) axis (Fig 8). Our findings suggest that heart failure- Transverse Aortic Constriction Leads to Depression-Like Behaviors induced inflammation and stress trigger CORT release via this axis, perturbing σ 1 R expression in brain, particularly in the hippocampus. Lower σ 1 R activity reduces mitochondrial ATP production, further impairing brain function and resulting in depression. Depression-like behaviors may also aggravate heart function by dysregulating sympathetic nerve activity [11].
Previous reports indicate that neuronal σ 1 R downregulation impairs IP 3 R-mediated mitochondrial Ca 2+ transport and ATP production in neurons [9]. ATP serves as not only an energy supply but also acts as an extracellular signaling molecule between neurons and glial cells [27]. For example, in hippocampal slices, norepinephrine treatment promotes glial cell ATP production and increases synaptic transmission efficacy [28]. SA4503 treatment enhances IP 3 -mediated Ca 2+ mobilization into mitochondria and mitochondrial ATP production in cardiomyocytes [8] and heart tissue in vivo (Fig 2F). Our present study reveals that SA4503 treatment antagonizes TAC-induced reduced σ 1 R expression in hippocampal astrocytes and enhances ATP production. Cao et al. reported that mice lacking the IP 3 receptor type 2 (Itpr2 -/-) display depression-like behaviors concomitant with reduced ATP levels in prefrontal cortex and hippocampus [29]. They also showed that acute increases in ATP ameliorate depression-like behavior by stimulating the P2X2 receptor. These authors conclude that reduced levels of hippocampal and cortical astrocyte-derived ATP are critical for induction of depression-like behaviors [29]. Combined with our findings, these studies indicate that reduced ATP production followed by σ 1 R downregulation in hippocampal astrocytes likely triggers TAC-induced depression-like behavior. Schematic representation of the hypothesis tested and results obtained. Shown is a model depicting mechanisms underlying TAC-induced depressive-like behavior, as suggested by findings presented here. TAC-induced cardiac σ 1 R downregulation induces cardiac dysfunction. At the same time, levels of peripheral pro-inflammatory cytokines and corticosterone increase, aggravating cardiac dysfunction and downregulating brain hippocampal σ 1 R levels. σ 1 R stimulation with SA4503 blocks TAC-induced σ 1 R downregulation in heart and in hippocampus. We also found that TAC-dependent σ 1 R downregulation in astrocytes occurs primarily in hippocampal CA1 and DG regions. Ito et al. reported that treatment with a different σ 1 R agonist, PRE084, antagonizes reduced σ 1 R expression seen in brain extracts after aortic banding in mice fed a high-salt diet [11]. We have reported that several σ 1 R agonists such as fluvoxamine, pentazocine and SA4503 restore reduced σ 1 R levels following heart failure in heart left ventricles [8,16,30]. Co-administration of σ 1 R antagonist, NE100, blocked those cardioprotective effects, whereas NE-100 administration did not block the elevated σ 1 R levels by SA4503 [8,16,30]. These results are consistent with the result in hippocampus of TAC mice in this report. Overall, improvement of depressive behabiors by SA4503 in TAC mice is mediated by σ 1 R stimulation rather than restoration of σ 1 R levels. Thus, mechanism underlying elevation of SA4503-induced σ 1 R levels remains unclear.
Unlike previous report [11], TAC with or without SA4503 treatment did not alter sympathetic nerve activity, as evaluated in our study by heart rate and blood pressure analysis. On the other hand, we found that plasma CORT levels increased at 4-and 6-week time points after TAC with concomitant CORT increases in the brain. Furthermore, chronic treatment of TAC mice with CORT significantly reduced hippocampal σ 1 R expression, especially in the CA1 region and DG (Fig 4), and mifepristone-mediated glucocorticoid receptor inhibition antagonized this effect in hippocampus (Fig 5). In support of our findings, glucocorticoid receptors are highly expressed in rat hippocampus [31]. In studies of human cohorts exhibiting heart failure, mortality increased significantly in patients with high plasma cortisol, a known risk factor for increased mortality by heart disease [32]. Taken together, increased plasma CORT levels are a risk factor for both reduced hippocampal σ 1 R expression and TAC-induced depressionlike behaviors. However, further studies are required to define additional factors underlying TAC-induced CORT production and CORT-induced σ 1 R downregulation.
Accumulating clinical evidence suggests that depression is a risk factor for cardiac disease and that SSRI treatment reduces post-MI morbidity and mortality following heart failure [3]. Hashimoto et al. recently suggested that σ 1 R function links cardiovascular disease to depression [33]. Indeed, mice subjected to aortic banding and infused with the σ 1 R agonist PRE084 showed suppressed sympathetic nerve activation with concomitant improvements in cardiac function [11]. However, as noted, our study suggests that SA4503 had no effect on sympathetic nerve activity. Previously we demonstrated that SA4503 directly stimulates σ 1 R in cardiomyocytes, rescuing reduced ATP levels and cardiac dysfunction [8]. Thus, although sympathetic nerve regulation by σ 1 R may account for some amelioration of cardiac dysfunction, our study suggests that glucocorticoid receptors likely have a critical function in TAC-induced depression-like behaviors. Interestingly, depressive patients exhibit increased secretions of peripheral stress hormones, including adrenal glucocorticoids, and treatment with anti-depressants including SSRIs normalize that secretion [1]. The plasma levels of cortisol and CORT increase in cardiovascular disease or depressive patients [25,34], and prescribed treatment with high-dose glucocorticoids is associated with increased risk for cardiovascular disease [35]. A previous report also suggested that inflammatory cytokine levels increase in cardiovascular disease or depressive patients [36,37]. Moreover, HPA axis is activated following stress with concomitant increases in inflammatory cytokine levels, and increased plasma CORT upregulate pro-inflammatory gene expression [38][39][40][41]. Dantzer et al. reported that peripheral lipopolysaccharide (LPS) administration induced both depression-like behavior and pro-inflammatory cytokine production in mice 24 hr later [42]. Relevant to these findings, we showed that brain CORT levels increased in parallel with peripheral CORT levels (1.11±0.11 ng/mg tissue (sham) vs. 1.35±0.16 ng/mg tissue (TAC 6 weeks), suggestive of an inflammatory response. Overall, we conclude that peripheral induction of pro-inflammatory cytokines by prolonged TAC activates the HPA axis, although further studies are required to define pro-inflammatory cytokines underlying these activities.
Depression-like behaviors induced by chronic CORT administration are often accompanied by concomitant mitochondrial dysfunction, including reduced mitochondrial DNA content and membrane potential [43]. In Flinders rats, a genetic model of depression, depression-like behaviors are associated with decreased numbers of mitochondria in hippocampal neurons [44]. Likewise, overall levels of proteins, DNA, and ATP are significantly reduced in depressive patients [45]. Importantly, σ 1 R, a MAM-associated protein, and ATP content markedly decreased in hippocampus following prolonged TAC, and depression-like behaviors are closely associated with mitochondrial dysfunction in the hippocampus. Here, we focused on σ 1 R as it functions as a chaperone for the IP 3 receptor, which is critical for ATP production following mitochondrial Ca 2+ transport [8,10,24]. Heart muscle consumes large amounts of ATP [46], and SA4503 enhances heart ATP production, thereby decreasing cardiac dysfunction. Moreover, other studies suggest that heart disease is a risk factor of neurological disease including depression or cognitive dysfunction [3,47].

Conclusions
Downregulation of hippocampal σ 1 R expression was induced by increasing plasma CORT levels following heart dysfunction. The TAC model proved useful to study mechanisms linking cardiovascular disease and depression. We also found that SA4503, a potent and selective σ 1 R agonist, ameliorates TAC-induced cardiac dysfunction and depression-like behaviors through σ 1 R upregulation. SA4503 has recently been tested as a treatment for depression in stroke patients in a phase II clinical trial (NCT00639249). Thus, SA4503 is an attractive candidate to prevent depression associated with cardiovascular disease.