Transcription Factor Foxo3a Prevents Apoptosis by Regulating Calcium through the Apoptosis Repressor with Caspase Recruitment Domain*

Background: Foxo3a plays a pivotal role in regulating apoptosis; however, its role in controlling cardiac apoptosis remains to be fully elucidated. Results: Foxo3a prevents cardiac apoptosis by regulating calcium through ARC. Conclusion: Foxo3a and ARC constitute an anti-apoptotic pathway that regulates calcium homeostasis in the heart. Significance: Our results provide important information for exploring the beneficial effects of this pathway on apoptosis-related cardiac diseases. Apoptosis can occur in the myocardium under a variety of pathological conditions, including myocardial infarction and heart failure. The forkhead family of transcription factor Foxo3a plays a pivotal role in apoptosis; however, its role in regulating cardiac apoptosis remains to be fully elucidated. We showed that enforced expression of Foxo3a inhibits cardiomyocyte apoptosis, whereas knockdown of endogenous Foxo3a sensitizes cardiomyocytes to undergo apoptosis. The apoptosis repressor with caspase recruitment domain (ARC) is a potent anti-apoptotic protein. Here, we demonstrate that it attenuates the release of calcium from the sarcoplasmic reticulum and inhibits calcium elevations in the cytoplasm and mitochondria provoked by oxidative stress in cardiomyocytes. Furthermore, Foxo3a is shown to maintain cytoplasmic and mitochondrial calcium homeostasis through ARC. We observed that Foxo3a knock-out mice exhibited enlarged myocardial infarction sizes upon ischemia/reperfusion, and ARC transgenic mice demonstrated reduced myocardial infarction and balanced calcium levels in mitochondria and sarcoplasmic reticulum. Moreover, we showed that Foxo3a activates ARC expression by directly binding to its promoter. This study reveals that Foxo3a maintains calcium homeostasis and inhibits cardiac apoptosis through trans-activation of the ARC promoter. These findings provided novel evidence that Foxo3a and ARC constitute an anti-apoptotic pathway that regulates calcium homeostasis in the heart.

hypertrophy and heart failure. Thus, it necessitates the identification of the molecules that are able to regulate cardiac apoptosis.
The forkhead family of transcription factors participates in regulating diverse cellular functions such as apoptosis, differentiation, metabolism, proliferation, and survival (3). Foxo3a can either induce or prevent apoptosis. For example, its activation in hematopoietic (4,5) and neuronal cells (6) results in the induction of apoptosis. In contrast, Foxo3a is necessary for the maintenance of neutrophil survival. Such a differential cellular response of Foxo3a activation can be related to the cell typespecific regulation of pro-and anti-apoptotic genes. Foxo3a is expressed in the heart and skeletal muscle (7)(8)(9). We and others have shown that Foxo3a can negatively regulate cardiac hypertrophy (8,10,11). It is not yet clear whether Foxo3a participates in the regulation of cardiac apoptosis.
Mitochondrial Ca 2ϩ homeostasis plays a critical role in maintaining cell survival. Its disruption by Ca 2ϩ overload can lead to apoptosis (12). For example, Ca 2ϩ overload promotes the opening of the mitochondrial permeability transition pore, and agents that block the mitochondrial permeability transition pore opening inhibit apoptosis (12). Oxidative stress induces a significant change in mitochondrial Ca 2ϩ flux leading to mitochondrial destabilization and apoptosis (13). Post-ischemic dopamine treatment of contractile dysfunction activates proapoptotic signal cascades via Ca 2ϩ -dependent mitochondrial damage (14). Sarcoplasmic reticulum (SR) 3 and mitochondria locate close to each other in cardiomyocytes (15). There is a tight coupling of Ca 2ϩ signaling between SR release sites and nearby mitochondria, because the focal SR Ca 2ϩ release results in Ca 2ϩ microdomains sufficient to promote local mitochondrial Ca 2ϩ uptake (16,17). Thus far, it remained unknown as to whether Foxo3a can regulate mitochondrial Ca 2ϩ homeostasis.
The heart has evolutionarily developed a highly expressed anti-apoptotic protein, ARC (18,19). It was originally identified to be a caspase-inhibiting protein and can specifically inhibit the activation of caspase-2 and -8, thereby blocking apoptosis induced by a variety of stimuli requiring the engagement of these caspases (18). Further studies revealed that ARC also may elicit its anti-apoptotic function by other means. It can interact with Fas, FADD, and Bax (20,21), inhibit cytochrome c release (22), and maintain mitochondrial membrane potential (23,24). Despite ARC's abundant expression and blocking apoptosis induced by either intrinsic or extrinsic stimulus, cardiomyocytes still undergo apoptosis under pathological conditions such as oxidative stress (23,25) and hypoxia (26 -28). Apoptosis is controlled by a complex interplay between pro-and antiapoptotic factors. The occurrence of apoptosis under pathological conditions indicates that this interplay is imbalanced. It remains to be further elucidated as to how ARC is dysregulated under the pathological conditions. Also, it is not yet clear whether ARC is involved in the maintenance of mitochondrial Ca 2ϩ homeostasis.
This work aimed to elucidate the role of Foxo3a in cardiac apoptosis. Our results show that Foxo3a can inhibit apoptosis induced by oxidative stress. Foxo3a knock-out mice exhibit accelerated myocardial infarction upon ischemia/ reperfusion. Furthermore, ARC can maintain calcium homeostasis, and Foxo3a regulates calcium through ARC. Finally, our results revealed that ARC is a transcription target of Foxo3a. Taken together, our study revealed a novel antiapoptotic pathway in which Foxo3a regulates calcium through transactivating ARC.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatment-Cardiomyocytes were isolated from 1-to 2-day-old Wistar rats as we described previously (10,29). In brief, after the dissected hearts were washed, they were minced in HEPES-buffered saline solution containing 130 mmol/liter NaCl, 3 mmol/liter KCl, 1 mmol/liter NaH 2 PO 4 , 4 mmol/liter glucose, and 20 mmol/liter HEPES (pH adjusted to 7.35 with NaOH). Tissues were then dispersed in a series of incubations at 37°C in HEPES-buffered saline solution containing 1.2 mg/ml pancreatin and 0.14 mg/ml collagenase (Worthington). After centrifugation, cells were resuspended in Dulbecco's modified Eagle's medium/F-12 (Invitrogen) containing 5% heat-inactivated horse serum, 0.1 mmol/liter ascorbate, insulin-transferring sodium selenite medium supplement, 100 units/ml penicillin, 100 g/ml streptomycin, and 0.1 mmol/liter bromodeoxyuridine. The dissociated cells were preplated at 37°C for 1 h. The cells were then diluted to 1 ϫ 10 6 cells/ml and plated in 10 g/ml laminin-coated separate culture dishes according to the specific experimental requirements. Neonatal mouse cardiomyocytes were cultured as described elsewhere (30,31). In brief, the hearts from wild type and Foxo3a knock-out mice were harvested, minced, and dispersed by 1.2 mg/ml pancreatin and 0.625 mg/ml collagenase (Worthington). Myocytes and nonmyocytes were sepa-rated by pre-plating for 1 h. Treatment of cells with hydrogen peroxide was carried out as we described previously (32). 10 M Ca 2ϩ chelator BAPTA-AM (Invitrogen) was administered 1 h before hydrogen peroxide or anoxia/ reoxygenation treatment. Anoxia/reoxygenation was performed as described elsewhere (33).
ARC Transgenic Mice-For creating the ARC transgenic mice, rat ARC coding sequence (GenBank TM accession number NM_053516) was cloned to the vector, p␣MHC-clone26 (kindly provided by Dr. Zhongzhou Yang), under the control of the ␣-myosin heavy chain promoter. Microinjection was performed following standard protocols. The primers for genotyping ARC transgenic mice include the following: forward primer in the ␣-MHC promoter, 5Ј-CACATAGAAGCCTAGCCC-ACA-3Ј, and the reverse primer in the ARC coding sequence, 5Ј-TTAGGTGTTCTCACAACCTTC-3Ј.
Genotyping of Foxo3a Knock-out Mice-Foxo3a knock-out (KO) mice were purchased from the Mutant Mouse Regional Resource Center. Foxo3a ϩ/Ϫ mice were interbred to give knock-out mice (Foxo3a Ϫ/Ϫ ), which were used for further studies. Mice were genotyped by multiplex PCR (primers and conditions are available from Mutant Mouse Regional Resource Center). All experiments were performed on Foxo3a Ϫ/Ϫ mice and their wild type littermates (Foxo3a ϩ/ϩ ) and were approved by government authorities.
Adenoviral Vector Construction and Infection-The adenoviruses harboring ARC, wild type Foxo3a (WTFoxo3a), and the constitutively active form of human Foxo3a (caFoxo3a) were as we described previously (10,34). The adenoviruses harboring rat Foxo3a and ARC RNAi were constructed using pSilencer TM adeno 1.0-CMV system. The rat Foxo3a RNAi target sequence is 5Ј-CAAGTACACCAAGAGCCGA-3Ј. A nonrelated and scrambled RNAi without any other match in the rat genomic sequence was used as a control (5Ј-TCAGACAGACAGACA-GACC-3Ј). The rat ARC RNAi target sequence is 5Ј-ACTGT-GAGCATGCCAGACC-3Ј, and the scrambled RNAi sequence is 5Ј-GTGCATCAGACTACCAGGC-3Ј. All viruses were amplified in HEK-293 cells. Cells were infected at the indicated multiplicity of infection (m.o.i.) for 60 min. After washing with phosphate-buffered saline (PBS), culture medium was added, and cells were cultured until the indicated time.
Preparations of Subcellular Fractionation-Mitochondrionenriched heavy membranes and cytosolic fractions were prepared as described previously (35). Briefly, cells were washed twice with PBS, and the pellet was suspended in 0.5 ml of buffer (20 mmol/liter HEPES, pH 7.5, 10 mmol/liter KCl, 1.5 mmol/ liter MgCl 2 , 1 mmol/liter EGTA, 1 mmol/liter EDTA, 1 mmol/ liter DTT, 0.1 mmol/liter PMSF, 10 mg/ml each of leupeptin, aprotinin, and pepstatin A) containing 250 mmol/liter sucrose. The cells were homogenized by 10 strokes in a Dounce homogenizer. The homogenates were centrifuged twice at 750 ϫ g for 5 min at 4°C. The supernatants were centrifuged at 10,000 ϫ g for 15 min at 4°C to collect mitochondrion-enriched heavy membranes. The final supernatants are referred to as cytosolic fractions. The isolation of SR-enriched membrane fractions were carried out as described previously (36,37). In brief, the cells were homogenized and centrifuged at 12,000 ϫ g for 10 Foxo3a Regulates Cardiac Apoptosis min. The supernatants were centrifuged at 100,000 ϫ g for 45 min. The pellets are referred to as SR fractions.
Luciferase activity assay was performed using the Dual-Luciferase reporter assay system (Promega) according to the manufacturer's instructions. Cells were lysed and assayed for luciferase activity 24 h after transfection. 20 l of protein extracts were analyzed in a luminometer. Firefly luciferase activities were normalized to Renilla luciferase activity.
Quantitative Real Time-PCR-Quantitative RT-PCR was performed as we described previously (32). In brief, total RNA was isolated using TRIzol (Invitrogen). RNA was reverse-transcribed using Oligo(dT) and amplified using a TaqMan assay kit (TOYOBO). The samples were run in triplicate using the Applied Biosystems 7000 sequence detector according to the manufacturer's instructions. The results were standardized to control values of GAPDH. The sequences of ARC primers were as follows: forward 5Ј-ATGGGTAACATGCAGGAG-CGC-3Ј and reverse 5Ј-GTCCAGCAGCAACCCAGAGTC-3Ј; GAPDH forward primer 5Ј-GCTAACATCAAATGGGGTG-ATGCTG-3Ј and reverse primer 5Ј-GAGATGATGACCCTT-TTGGCCCCAC-3Ј. PCR was run under the following conditions: 95°C for 60 s for 1 cycle; 95°C for 15 s, 55°C for 15 s, and 72°C for 45 s for 40 cycles. The specificity of the PCR amplification was confirmed by agarose gel electrophoresis.
Measurement of [Ca 2ϩ ] c -Cytosolic calcium was measured as described previously (40 -42). In brief, the cardiomyocytes cultured on a glass coverslip were loaded with 10 M Fluo-3-AM (Molecular Probes), in the HEPES solution without Ca 2ϩ at 37°C for 30 min, then washed twice with dye-free HEPES solution, and placed in a chamber on the stage of a laser scanning confocal microscope (Zeiss LSM510). Fluo-3 in cells was excited with light at 488 nm, and emitted fluorescence was detected at Ͼ500 nm. For fluorescence analysis, the regions of interest identified in the average fluorescence intensities over 20 regions of interest minus background were calculated for each frame and normalized for comparative purposes. The [Ca 2ϩ ] c values were calculated by Mn 2ϩ quenching.
Detections of [Ca 2ϩ ] m -The mitochondrial calcium indicator Rhod-2 AM (Molecular Probes) was used to measure changes of [Ca 2ϩ ] m as described previously (43,44). Briefly, cardiomyocytes were loaded with 5 M Rhod-2 AM for 2 h at 4°C and further incubated for 2 h at 37°C in the culture medium. This two-step cold loading/warm incubation protocol achieves exclusive loading of Rhod-2 into the mitochondria. The Rhod-2 images were captured with excitation at 543 nm and detected at Ͼ560 nm. Rhod-2 fluorescence intensities (F) in each experiment were normalized to the average base-line fluorescence for the same region (F 0 ).
Detections of [Ca 2ϩ ] SR -The low affinity calcium indicator, Fluo-5N-AM (Molecular Probes), was used to assess changes in the intraluminal SR Ca 2ϩ concentration as described previously (45,46). In brief, the cardiomyocytes were loaded with 5 M Fluo-5N-AM at 37°C for 2 h, rinsed, and incubated in culture medium for a further 1.5 h at 37°C to allow de-esterification and outward leak of the cytosolic indicator. Fluo-5N was excited with light at 488 nm, and the emitted fluorescence was detected at Ͼ500 nm. Fluo-5N fluorescence intensities (F) in each experiment were normalized to the average base-line fluorescence for the same region (F 0 ).
Detection of Mitochondrial Membrane Potential (⌬⌿ m )-⌬⌿ m was measured using tetramethylrhodamine methyl ester (TMRM) as described previously (39). In brief, the cells were Foxo3a Regulates Cardiac Apoptosis MARCH 22, 2013 • VOLUME 288 • NUMBER 12 JOURNAL OF BIOLOGICAL CHEMISTRY 8493 stained with 10 nM TMRM (Molecular Probes) at 37°C for 30 min. TMRM images were captured using a laser scanning confocal microscope (Zeiss LSM510) with excitation at 543 nm and emission at Ͼ560 nm.

Ischemia/Reperfusion (I/R), Hemodynamic Assessment, LDH
Release Assay, and Analysis of Infarction Sizes-Mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (5 mg/kg) and fixed in the supine position, and tracheot-

Foxo3a Regulates Cardiac Apoptosis
omy was performed to provide artificial ventilation (0.2-ml tidal volume, 110 breaths/min) with a rodent ventilator supplemented with 100% oxygen. The left anterior descending coronary artery was identified and ligated with a slipknot using 8-0 silk suture at the inferior border of the left auricle. Myocardial ischemia was confirmed by the obvious cyanotic appearance of the left ventricle and S-T segment elevation on the electrocardiogram. After 30 min of ischemia, the knot was released to permit reperfusion of the heart, which was also confirmed by obvious S-T segment change. After removal of air and blood, the chest was closed, and the animal was removed from the respirator and transferred back to its cage. Sham-operated mice were prepared identically without undergoing the occlusion of the silk suture. 24 h after reperfusion, the mice were anesthetized as described above, and the LV hemodynamic measurements were conducted by an experienced investigator, who was blind to the treatment, through a closed-chest catheterization. A 1.4-F microtipped catheter (SPR-839, Millar Instruments) was inserted into the right carotid artery and advanced into the left ventricle. Hemodynamic parameters such as left ventricular end-diastolic pressure, left ventricular maximum first derivative of pressure (LV dP/dt max ), minimum first derivative of pressure (LV dP/dt min ), and LV ejection fraction were computed. The heparin-blood at the end of the hemodynamic assessment was collected from all groups, and the concentration of LDH in the plasma was assayed using an LDH ELISA kit (Adlitteram Diagnostic Laboratories). To determine myocardial infarct sizes, the thoracotomy was reopened; the suture was reoccluded, and 2% of Evans blue (Sigma) was injected into the left ventricular cavity to delineate the ischemic zone from the nonischemia zone. The heart was immediately removed and flushed with ice-cold saline and frozen in a Ϫ80°C freezer. Each heart was then horizontally cut into five slices that were incubated in 1.0% 2,3,5-triphenyltetrazolium chloride (Sigma) for 15 min at 37°C for demarcation of the viable and nonviable myocardium within AAR. Infarct myocardium appears yellowish   Evaluation of Apoptosis-To determine apoptosis in the heart sections, we used In Situ Cell Death Detection Kit, Fluorescein (Roche Diagnostics). The detections were performed as we described elsewhere (32,47).
Detection of Caspase-3 Activity-Caspase-3 activity was detected using an assay kit (R&D Systems). The assay procedures were according to the kit instructions.
Statistical Analysis-Paired data were evaluated by Student's t test. A one-way analysis of variance was used for multiple comparisons. A p value of Ͻ0.05 was considered significant.

Foxo3a Can Suppress Cytosolic Ca 2ϩ
Elevations and Apoptosis-Reactive oxygen species play an important role in mediating apoptosis in cardiomyocytes. In this study, we observed that hydrogen peroxide or anoxia/reoxygenation treatment leads to the increase of cytosolic Ca 2ϩ levels [Ca 2ϩ ] c in cardiomyocytes (Fig. 1A). Administration of the Ca 2ϩ chelator, BAPTA-AM, inhibits apoptosis (Fig. 1B), suggesting that Ca 2ϩ is necessary for the initiation of apoptosis.
Foxo3a can trigger or inhibit apoptosis depending on the cell types or cellular context (6,48). We tested whether Foxo3a can influence apoptosis in cardiomyocytes. Enforced expression of the constitutively active form of Foxo3a (caFoxo3a) itself did not induce apoptosis; however, it could prevent apoptosis triggered by hydrogen peroxide (Fig. 1C). Because Ca 2ϩ is necessary for the initiation of apoptosis as shown in Fig. 1B (Fig. 1D). We then tested whether endogenous  (Fig. 1E). Concomitantly, a significant amount of cells underwent apoptosis (Fig. 1F). Furthermore, we isolated cardiomyocytes from Foxo3a knock-out mice and observed a higher level of [Ca 2ϩ ] c in these cells upon treatment with hydrogen peroxide or anoxia/reoxygenation, and noticeably, re-expression of caFoxo3a attenuated [Ca 2ϩ ] c levels (Fig.  1G). Hydrogen peroxide or anoxia/reoxygenation induced more cardiomyocytes from Foxo3a knock-out mice to undergo apoptosis, and this can be inhibited by re-expression of caFoxo3a (Fig. 1H). Finally, we further characterized whether apoptosis indeed occurred under our experimental condition. Enforced expression of caFoxo3a suppressed the activation of caspase-3 in neonatal rat cardiomyocytes (Fig. 2, A and B) or cardiomyocytes from Foxo3a-deficient mice (Fig. 2C). Pretreatment of the cells with caspase inhibitor Z-VAD-fmk suppressed both apoptosis (Fig. 2, D and E) and caspase-3 activation (Fig. 2, F and G) triggered by hydrogen peroxide or anoxia/ reoxygenation. Taken together, these data indicate that Foxo3a is able to suppress [Ca 2ϩ ] c elevations and apoptosis.
Foxo3a Knock-out Mice Exhibit an Enlarged Myocardial Infarction Size-Apoptosis is a kind of death form in myocardial infarction. We tested whether Foxo3a plays a role in regulating myocardial infarction. Foxo3a deficient mice exhibited a larger infarction size upon I/R (Fig. 3A), and a more severe injury in cardiac function (Fig. 3B). We analyzed apoptosis and observed more apoptotic cells in Foxo3a deficient mice upon I/R (Fig. 3C).
These results suggest that Foxo3a participates in inhibiting apoptosis and myocardial infarction in the animal model. Furthermore, our data shows that LDH release was significantly increased in hearts from Foxo3a deficient mice (Fig. 3D), which suggests that loss of Foxo3a exacerbated myocardial injury induced by I/R. ARC Attenuates Cytosolic Ca 2ϩ Elevations-To understand the relationship between ARC and Ca 2ϩ in the apoptotic program of cardiomyocytes, we tested whether ARC could influence [Ca 2ϩ ] c . ARC levels were decreased upon hydrogen peroxide treatment (Fig. 4A). Knockdown of endogenous ARC could sensitize hydrogen peroxide to induce [Ca 2ϩ ] c elevations (Fig. 4B), apoptosis (Fig. 4C), and the activation of caspase-3 (Fig. 4D). Enforced expression of ARC attenuated [Ca 2ϩ ] c elevations induced by hydrogen peroxide (Fig. 4E). Concomitantly, apoptosis (Fig. 4F) and caspase-3 activation could be inhibited (Fig. 4G). Furthermore, we isolated the cardiomyocytes from ARC transgenic mice and observed that the [Ca 2ϩ ] c increases evoked by stimulation with hydrogen peroxide were significantly less than that from wild type mice (Fig. 4H). Thus, it appears that ARC is able to attenuate cytoplasmic calcium elevation induced by hydrogen peroxide in cardiomyocytes.
ARC Participates in the Control of Ca 2ϩ Release from Sarcoplasmic Reticulum-How can ARC lead to a reduction of [Ca 2ϩ ] c ? Sarcoplasmic reticulum release of Ca 2ϩ has been proved to contribute to [Ca 2ϩ ] c elevations during apoptosis. We tested whether ARC can influence sarcoplasmic reticulum release of Ca 2ϩ . ARC has been shown previously to be localized in the cytoplasm and mitochondria (20,35), but it is not yet clear whether ARC is distributed in the sarcoplasmic reticulum. We analyzed ARC distributions in the subcellular organelles and observed that a portion of ARC was distributed in the sar- coplasmic reticulum as analyzed by immunoblotting and immunofluorescence (Fig. 5A). Subsequently, we investigated whether ARC can influence Ca 2ϩ within the sarcoplasmic reticulum ([Ca 2ϩ ] SR ). Hydrogen peroxide induced a reduction of [Ca 2ϩ ] SR . Enforced expression of ARC elevated [Ca 2ϩ ] SR upon treatment with hydrogen peroxide (Fig. 5B). We attempted to understand whether endogenous ARC plays a role in regulating [Ca 2ϩ ] SR , and we observed that knockdown of ARC could sensitize the reduction of [Ca 2ϩ ] SR upon treatment with 10 M hydrogen peroxide (Fig. 5C). These data indicate that ARC is able to regulate sarcoplasmic reticulum Ca 2ϩ release into the cytoplasm.
ARC Attenuates Mitochondrial Ca 2ϩ Uptake-It has been well documented that the imbalance in mitochondrial Ca 2ϩ ([Ca 2ϩ ] m ) leads to apoptosis. We tested whether ARC was involved in the maintenance of [Ca 2ϩ ] m homeostasis. Enforced expression of ARC reduced [Ca 2ϩ ] m rises upon treatment with hydrogen peroxide (Fig. 6A). Knockdown of ARC sensitized the elevation of [Ca 2ϩ ] m upon treatment with 10 M hydrogen peroxide (Fig. 6B). We isolated cardiomyocytes from ARC transgenic mice, and they exhibited a lower level of [Ca 2ϩ ] m than those from the wild type mice upon hydrogen peroxide treatment (Fig. 6C), suggesting that ARC serves to inhibit the elevation of [Ca 2ϩ ] m .
To further understand the role of ARC in regulating [Ca 2ϩ ] m , we made a construct in which ARC was fused to mitochondrion-targeted GFP, and we termed this construct as M/GFP/ ARC, which was specifically localized in mitochondria (Fig.  6D). The empty mitochondrion-targeted GFP was termed as M/GFP. M/GFP/ARC was able to attenuate [Ca 2ϩ ] m elevations induced by hydrogen peroxide (Fig. 6E). Also, it could inhibit the collapse of mitochondrial membrane potential (Fig. 6F). We analyzed cytochrome c distributions in cells expressing M/GFP/ARC. In the control cells without treatment, cytochrome c distribution pattern was coincident with that of M/GFP. Upon treatment with hydrogen peroxide, cytochrome c and M/GFP showed a differential pattern. However, in cells expressing M/GFP/ARC, the cytochrome c distribution pattern was not significantly altered (Fig. 6G).
Foxo3a Regulates Ca 2ϩ through ARC-Foxo3a is a transcription factor. How can it influence Ca 2ϩ machinery in cardiomyocytes? Because of the ability of ARC to regulate Ca 2ϩ , we tested whether ARC can be a downstream mediator of Foxo3a to target Ca 2ϩ . Knockdown of ARC could abolish the effect of Foxo3a on attenuating [Ca 2ϩ ] c elevations (Fig. 7A) and apoptosis (Fig. 7B) induced by hydrogen peroxide. The effects of Foxo3a on attenuating caspase-3 activation (Fig. 7C), [Ca 2ϩ ] m increases (Fig. 7D), and cytochrome c release (Fig. 7E) were also abolished upon ARC knockdown. These data indicate that ARC is a downstream mediator of Foxo3a in regulating calcium homeostasis.
ARC Is a Transcriptional Target of Foxo3a-We explored the relationship between ARC and Foxo3a. An elevated level of ARC mRNA in caFoxo3a-expressed cardiomyocytes could be observed (Fig. 8A). Also, ARC protein levels were elevated upon caFoxo3a stimulation (Fig. 8B). We further tested whether endogenous Foxo3a participates in the regulation of ARC. Knockdown of endogenous Foxo3a by RNAi led to a reduction in mRNA (Fig. 8C) and protein (Fig. 8D) levels of ARC.
The influence of Foxo3a on ARC expression led us to consider whether ARC is a transcriptional target of Foxo3a. To address this consideration, we analyzed the rat ARC promoter region, and we found that the promoter region of ARC contains three potential Foxo3a-binding sites (Fig. 8E). We first tested whether Foxo3a can regulate ARC promoter activity. Luciferase assay revealed that although enforced expression of wild type Foxo3a (WTFoxo3a) and caFoxo3a could stimulate ARC promoter activity, caFoxo3a had a stronger effect. Furthermore, only the full-length ARC promoter could be activated by Foxo3a, indicating that the BS1 was responsible for the luciferase activity (Fig. 8F). Knockdown of endogenous Foxo3a by RNAi significantly reduced ARC promoter activity in car-diomyocytes (Fig. 8G). To determine Foxo3a-binding sites in the ARC promoter in vivo, we performed ChIP assays using primers directed against the BS1, BS2, and BS3, respectively. Foxo3a bound to the BS1 but not BS2 and BS3 in the ARC promoter (Fig. 8H). Finally, we analyzed ARC levels in Foxo3a knock-out mice and observed that these mice had a low level of ARC in comparison with the wild type mice (Fig. 8I). Taken together, we demonstrate that ARC is a transcriptional target of Foxo3a.

DISCUSSION
Although Foxo3a plays a role in regulation of apoptosis in a variety of cell types, its role in cardiac apoptosis remains to be fully understood. This work demonstrated that Foxo3a inhibits  FIGURE 8. ARC is a transcriptional target of Foxo3a. A, Foxo3a stimulates ARC mRNA expression. Neonatal rat cardiomyocytes were infected with the adenoviral caFoxo3a or ␤-gal at an m.o.i. of 80. Cells were harvested for the analysis of ARC mRNA by quantitative RT-PCR. The values were normalized to that of GAPDH. *, Ͻ 0.05 versus control. B, Foxo3a up-regulates ARC protein levels. Neonatal rat cardiomyocytes were infected as described for A. Cells were harvested for the analysis of ARC protein by immunoblotting. The blots shown here are the representative blots from three independent experiments. Numbers above immunoblots show the ratios of the band intensity of ARC to that of actin. C, knockdown of Foxo3a leads to a reduction in ARC mRNA levels. Neonatal rat cardiomyocytes were infected with adenoviral Foxo3a-RNAi or its scrambled form (Foxo3a-S-RNAi) at an m.o.i. of 100. Cells were harvested 48 h after infection for the analysis of ARC mRNA levels by quantitative RT-PCR, *, p Ͻ 0.05 versus control. D, knockdown of Foxo3a leads to a reduction in ARC protein levels. Neonatal rat cardiomyocytes were infected as described for C. Cells were harvested 48 h after infection for the analysis of ARC protein levels by immunoblotting. The blots shown here are the representative blots from three independent experiments. Numbers above immunoblots show the ratios of the band intensity of ARC to that of actin. E, ARC promoter region has three potential Foxo3a-binding sites. Rat ARC promoter contains three potential Foxo3abinding sites (BS) indicated as BS1, BS2, and BS3. Three fragments of ARC promoter were synthesized and linked to luciferase reporter vector, respectively. F, Foxo3a stimulates the activity of ARC promoter containing BS1. HEK-293 cells were infected with adenoviral ␤-gal, wild type Foxo3a (wtFoxo3a), or caFoxo3a at an m.o.i. of 80. 24 h after infection, cells were transfected with the constructs of the empty vector (pGL-4) or ARC promoter constructs, respectively. Firefly luciferase activities were normalized to Renilla luciferase activities. G, inhibition of endogenous Foxo3a leads to a reduction of ARC promoter activity. Neonatal rat cardiomyocytes were infected with adenoviral Foxo3a-RNAi or its scrambled form (Foxo3a-S-RNAi) at an m.o.i. of 100. 24 h after infection, cells were transfected with the constructs of the empty vector (pGL-4) or the pGL-4-ARC promoter-1, respectively. Firefly luciferase activities were normalized to Renilla luciferase activities. *, p Ͻ 0.05 versus pGL-4-ARC promoter-1 alone. H, Foxo3a directly binds to BS1 of ARC promoter as analyzed by ChIP assay. Neonatal rat cardiomyocytes were infected with adenoviral ␤-gal, WTFoxo3a, or caFoxo3a at an m.o.i. of 80. Cells were harvested 24 h after infection for ChIP analysis. Chromatin-bound DNA was immunoprecipitated with the anti-Foxo3a antibody. The anti-actin antibody was used as a negative control (Neg). Immunoprecipitated DNA was analyzed by PCR using a primer combination that encompassed the three potential Foxo3a-binding sites in ARC promoter (BS1, BS2, and BS3), respectively. M, means marker/ladder. I, ARC protein levels in hearts from Foxo3a knock-out mice were less than that from wild type mice. The hearts from Foxo3a knock-out mice and their wild type littermates were harvested, and ARC levels were analyzed by immunoblotting. Data are expressed as mean Ϯ S.E. from three independent experiments. MARCH 22, 2013 • VOLUME 288 • NUMBER 12 apoptosis in cardiomyocytes, and the cardiac anti-apoptotic protein ARC attenuates the release of calcium from the sarcoplasmic reticulum and inhibits calcium elevations in the cytoplasm and mitochondria under oxidative stress. Foxo3a is shown to maintain calcium homeostasis and inhibit apoptosis in cardiomyocytes through ARC. Furthermore, ARC is a direct transcriptional target of Foxo3a. Thus, our results provide novel evidence that Foxo3a regulates cardiac apoptosis by maintaining calcium homeostasis through ARC.

Foxo3a Regulates Cardiac Apoptosis
The role of Foxo3a in apoptosis is dependent on the cell types and the cellular contexts. It can provoke apoptosis in hematopoietic (4,5) and neuronal cells (6) but can prevent apoptosis in neutrophils. This discrepancy is probably due to the multiple targets of Foxo3a, and its final transcriptional output is determined by the equilibrium of the pro-and anti-apoptotic factors. The identified downstream targets of Foxo3a include Fas ligand (FasL) (6), Bim (49), Mn-superoxide dismutase (50), and catalase (51). FasL activates the extrinsic apoptotic pathway by associating with Fas and consequently leading to the formation of the death-inducing signaling complex and caspase-8 activation. Bim is a member of the Bcl-2 family and counteracts Bcl-2 and Bcl-xL, thereby inducing apoptosis (52). Mn-superoxide dismutase and catalase can scavenge reactive oxygen species that are important apoptotic stimuli. This work revealed that Foxo3a is able to prevent apoptosis in cardiomyocytes. Notably, we have found that ARC is a transcriptional target of Foxo3a and necessary for Foxo3a to exert its anti-apoptotic function.
It has been shown that ARC is a calcium-binding protein and can suppress Ca 2ϩ -mediated apoptosis (53). This study demonstrated that ARC is present in the sarcoplasmic reticulum and participates in the regulation of mitochondrial Ca 2ϩ homeostasis by attenuating the release of [Ca 2ϩ ] SR induced by oxidative stress. ARC requires the C terminus to bind to the calcium, and the deletion of the C terminus leads to the inability of ARC to bind to calcium (53). Intriguingly, the subcellular localization of ARC is controlled by the C terminus. The mutation in threonine 149 in the C terminus results in the alteration of ARC subcellular localization and loss of ARC anti-apoptotic function (54). Thus, it appears that the C terminus is important for ARC function. Our data shed new light on understanding the novel molecular mechanism by which ARC inhibits apoptosis.
SR and mitochondria locate close to each other in cardiomyocytes (15). There is a tight coupling of Ca 2ϩ signaling between SR release sites and nearby mitochondria, because the SR Ca 2ϩ release results in Ca 2ϩ microdomains sufficient to promote local mitochondrial Ca 2ϩ uptake (16,17). In this study, we have shown that the increase in [Ca 2ϩ ] m may result from the reduction of [Ca 2ϩ ] SR . The membrane-permeable Ca 2ϩ chelator BAPTA-AM is highly selective for Ca 2ϩ and can be used to control the level of intracellular Ca 2ϩ . The efficiency of mitochondrial Ca 2ϩ uptake depends on the upstroke velocity of cytosolic Ca 2ϩ transients (55). It would be interesting to test whether BAPTA-AM can affect the effects of ARC on regulating calcium homeostasis in future studies.
ARC is involved in the control of myocardial infarction. ARC transgenic mice exhibit less myocardial infarction sizes (56), whereas ARC knock-out mice demonstrate accelerated myocardial infarction (47). ARC levels are significantly decreased in cardiomyocytes upon treatment with hydrogen peroxide and hypoxia (20,23). Furthermore, ARC levels are reduced upon heart failure (47) or ischemia (57). The expression reduction of a protein can be due to its decrease in synthesis and/or increase in degradation. ARC degradation is up-regulated upon apoptotic stimulation (57). This work has revealed that Foxo3a can transactivate ARC, and the reduced expression of Foxo3a contributes to ARC down-regulation. Thus, the dysregulation of ARC expression can be related to Foxo3a.
To keep the heart intact in both structure and function, it is necessary to prevent apoptosis so that the heart does not lose cardiomyocytes. Therefore, the development of anti-apoptotic strategies may prove useful as a means to prevent apoptosisrelated cardiac diseases leading to heart failure. This work provides novel evidence that Foxo3a and ARC constitute an antiapoptotic pathway that participates in the maintenance of calcium homeostasis. Our results can provide important information for exploring the beneficial effects of this pathway on apoptosis-related cardiac diseases.