Reactive Oxygen Species Suppress Cardiac NaV1.5 Expression through Foxo1

NaV1.5 is a cardiac voltage-gated Na+ channel αsubunit and is encoded by the SCN5a gene. The activity of this channel determines cardiac depolarization and electrical conduction. Channel defects, including mutations and decrease of channel protein levels, have been linked to the development of cardiac arrhythmias. The molecular mechanisms underlying the regulation of NaV1.5 expression are largely unknown. Forkhead box O (Foxo) proteins are transcriptional factors that bind the consensus DNA sequences in their target gene promoters and regulate the expression of these genes. Comparative analysis revealed conserved DNA sequences, 5′-CAAAACA-3′ (insulin responsive element, IRE), in rat, mouse and human SCN5a promoters with the latter two containing two overlapping Foxo protein binding IREs, 5′-CAAAACAAAACA-3′. This finding led us to hypothesize that Foxo1 regulates NaV1.5 expression by directly binding the SCN5a promoter and affecting its transcriptional activity. In the present study, we determined whether Foxo1 regulates NaV1.5 expression at the transcriptional level and also defined the role of Foxo1 in hydrogen peroxide (H2O2)-mediated NaV1.5 suppression in HL-1 cardiomyocytes using chromatin immunoprecipitation (ChIP), constitutively nuclear Foxo1 expression, and RNAi Foxo1 knockdown as well as whole cell voltage-clamp recordings. ChIP with anti-Foxo1 antibody and follow-up semi-quantitative PCR with primers flanking Foxo1 binding sites in the proximal SCN5a promoter region clearly demonstrated enrichment of DNA, confirming Foxo1 recruitment to this consensus sequence. Foxo1 mutant (T24A/S319A-GFP, Foxo1-AA-GFP) was retained in nuclei, leading to a decrease of NaV1.5 expression and Na+ current, while silencing of Foxo1 expression by RNAi resulted in the augmentation of NaV1.5 expression. H2O2 significantly reduced NaV1.5 expression by promoting Foxo1 nuclear localization and this reduction was prevented by RNAi silencing Foxo1 expression. These studies indicate that Foxo1 negatively regulates NaV1.5 expression in cardiomyocytes and reactive oxygen species suppress NaV1.5 expression through Foxo1.


Introduction
Alterations of ion channel activity play an important role in the development of cardiac arrhythmias. Na V 1.5 encoded by the SCN5a gene is a subunit of the voltage-gated cardiac Na + channel, and this channel activity determines cardiac excitability and conduction. Reactive oxygen species (ROS) are generated during myocardial infarction [1][2][3][4] and contribute to the development of life-threatening ventricular arrhythmias [5,6]. It has been shown that alteration of cardiac Na + channel activity immediately following ROS challenge requires activation of Ca 2+ -dependent calmodulin II kinase (CaMKII) [7]. CaMKII regulates Na + channel gating by phosphorylating Na V 1.5 to increase late Na + current, leading to the prolongation of the action potential duration and development of early afterdepolarization [8]. H 2 O 2chronically mediated downregulation of Na + channel activity requires activation of NF-KappaB to decrease Na V 1.5 expression by inhibiting SCN5a promoter activity [9]. Mutation of the NF-KappaB binding site in the SCN5a promoter eliminates 42.3% decrease of this promoter activity by H 2 O 2 in H9C2 myoblasts [9]. These findings indirectly support the notion that other signaling pathways are also significantly involved in the regulation of Na V 1.5 expression.
Foxo transcription factors form a subclass of the large family of Forkhead proteins characterized by the presence of a ''wingedhelix'' DNA-binding domain called Forkhead box O, which gave the name Foxo proteins to the family [10]. Post-translational modifications such as phosphorylation, acetylation and methyla-tion affect Foxo subcellular localization and transcriptional activity [11][12][13]. Foxo proteins harbor conserved threonine and serines that can be phosphorylated by different kinases, leading to their nuclear retention or exclusion [11]. Akt/protein kinase B (PKB) [14,15], serum and glucocorticoid-induced kinase 1 (SGK1) [16], cyclin-dependent kinase 2 (CDK2) [17] and IkB kinase (IKK) [18] phosphorylate Foxo proteins, resulting in their nuclear exclusion while c-Jun-N-terminal kinase (JNK) [19] and mammalian sterile 20-like kinase 1 (MST1) [20] phosphorylate Foxo proteins, causing their nuclear retention. Foxo proteins also contain acetylated lysines that can be deacetylated by a NAD-dependent histone deacetylase, silent mating type information regulation 2 homolog (Sirt1), leading to their nuclear retention [12]. Protein arginine methyltransferase 1 (PRMT1) methylates Foxo1 primarily at Arg248 and Arg250 within the consensus Akt phosphorylation motif. These methylations directly abrogate Akt-induced phosphorylation of Foxo1 at Ser253, which in turn results in increased Foxo1 nuclear localization [13]. Accumulating evidence suggests that Foxo1, 3 and 4 are expressed in the heart and they are critical in maintaining cardiac function and mediating cardiac stresses. In the heart, the Foxo family is believed to be involved in diverse activities, including response to oxidative stress [21][22][23], regulation of metabolism [24], modulation of KATP channels [25], promotion of autophagy [26], cell cycle control [27,28], and commitment to apoptosis [29]. In the nucleus, Foxo proteins target a conserved DNA binding sequence, 59-GTAAA(C/T)A-39 [30] or insulin response element (IRE), 59-CAAAA(C/T)A-39 [30][31][32] in the promoters of their target genes to affect these gene expression. The end result of Foxo transcriptional activity is generally associated with counteracting oxidative stress and promoting cell cycle arrest and apoptosis [11]. Silico promoter analysis on a list of 354 of the most differentially expressed genes in the thymocyte lineage predicted that the promoter of SCN5a has Foxo binding elements [33].
In this study, we tested the hypothesis that H 2 O 2 suppressing Na V 1.5 requires the Foxo1 signaling pathway. Our results showed that Foxo1 negatively regulated Na V 1.5 expression by directly binding the SCN5a promoter, and H 2 O 2 -mediated inhibition of Na V 1.5 expression required Foxo1.

Ethics
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the University Committee on Animal Resources (UCAR) of University of Rochester Medical Center (UCAR # 2007-060).

Cell Culture
HeLa cells (obtained from American Type Culture Collection) were maintained in 10 cm plates containing Dulbecco's modified Eagle's medium (high glucose) and 10% fetal bovine serum with 100 U/ml penicillin and 100 mg/ml streptomycin. A cardiomyocyte cell line, designated HL-1, from the AT-1 mouse atrial cardiomyocyte tumor lineage, was kindly provided by Dr. William C. Claycomb at Louisiana State University Health Sciences Center, New Orleans, USA. These cells were cultured as described previously [34]. The cells were cultured in Claycomb medium (JRH Biosciences, KS) supplemented with 10% fetal bovine serum (JRH Biosciences, KS), 2 mM L-glutamine, 100 mM norepinephrine, 100 U/ml penicillin and 100 mg/ml streptomycin in 10 cm plates pre-coated with fibronectin (BD Biosciences, PA). The medium was changed every 24-48 h. Mouse ventricular myocytes isolated from 10 week old FVB mice as previously described [35] were directly subjected to a chromatin immunoprecipitation (ChIP) assay.

Bioinformatics
The SCN5a genomic promoter sequences from mouse, rat, and human were downloaded from the University of California Santa Cruz genome browser [36] and comparative genomic analysis was done with the visualization tools for alignments (VISTA) algorithm (available from the VISTA web site). The Foxo protein biding sequences were compared.

Transfection and Luciferase Assays
Lipofectamine TM and plus TM reagents (Invitrogen) were used to transfect SNC5a reporters (SCN5a promoter-Luc plasmid [37] was generously provided by Dr. Hideko Kasahara in University of Florida College of Medicine, Florida, USA) and Foxo1 in the HeLa cells. Cells were seeded in 24-well plates and grown to 90% confluence. Transfections were performed according to the manufacturer's instructions. A Renilla reporter gene (Promega) was included as an internal control. Cell lysates were prepared for the luciferase assay 24 hours after transfection as described by the manufacturer (Promega). Data were analyzed and expressed as the normalized-fold changes over controls.

RT-PCR
Total RNA was extracted from HL-1 and HeLa cells by RNeasy Mini kit (QIAGEN) and reversely transcribed to cDNA with a high capacity cDNA reverse transcription kit (AB Applied Biosystems). Semi-quantitative PCR was conducted to measure mRNA expression levels across different experimental conditions as indicated under ''Results''. The primers used to amplify mouse(m) Na V 1.5, mß-actin, human(h) GAPDH, mGAPDH, hFoxo1, and mTR4 were summarized in table 1.

Western Blotting
HL-1 cells were rinsed with phosphate-buffered saline (PBS) twice, and protein was extracted in cold cell lysis buffer (Cell Signaling). Protein concentration was determined by a detergentcompatible protein assay (Bio-Rad). Equal amounts of protein were resolved by SDS-PAGE, transferred onto nitrocellulose membranes, blocked with 5% non-fat milk for 1 hour, and then incubated with the indicated primary antibody overnight at 4uC. After 1 hour incubation with the appropriate secondary antibody, specific signals were revealed by enhanced chemiluminescence reagent (Pierce). The primary antibodies used were as follows: rabbit anti-Foxo1 polyclonal (Cell Signaling, 1:1000), rabbit anti-Na V 1.5 polyclonal (Alomone Labs, 1:500), rabbit anti-GAPDH polyclonal (Millipore, 1:4000), and mouse anti-tubulin monoclonal (Sigma, 1:5000).

Fluorescen and Confocal Microscopy
HL-1 cells were dispersed on the glass in the center of the 32 mm dishes and grown to 80% confluence. HL-1 cells expressing GFP and Foxo1-GFP were fixed with 4% paraformaldehyde. Nuclear profiles of were revealed with a brief incubation in DAPI (Molecular Probes) prior to microscopic observation. Fluorescence was visualized with an inverted Olympus IX81 fluorescence confocal microscope and photographed for direct import into Image J. All images were processed in an identical manner to faithfully capture the real time images of each sample.

ChIP Assay
ChIP was used to study the interaction between Foxo1 and SCN5a promoter DNA in HL-1 cells, as previously described [38]. Approximately 1610 7 HL-1 cells were subjected to the ChIP assay using anti-Foxo1 antibody or rabbit control IgG as a control and the ChIP assay kit (Upstate Biotechnology), as previous described [38]. The immunoprecipitates were analyzed by immunoblot analysis using goat anti-Foxo1 (FKHR-C20; Santa Cruz Biotechnology Inc.), and by PCR assay to detect coimmunoprecipitated DNA using two pairs of SCN5a promoter-specific primers (forward 59-TGGTACATACCGTTTCAGGAC-39 and reverse 59 GCA-CACACACTCACACATAC -39; forward 59-TGGTACATA-CCGTTTCAGGAC-39 and reverse 59-ACGCATGCTCACG-CACACA -39) that flank the consensus Foxo1 binding sites (21484 to 21473 nt) in the mouse SCN5a promoter and a pair of primers (forward 59-GCCTCATTCGTTGCAGTGTGCC-39 and reverse 59-GCCAGTGCCTTGTGTGGACTCT-39) that flanked the DNA sequence (24453 to 24686) lacking Foxo1 biding site as a control.

Statistical Analysis
All values are expressed as Mean6SEM. Data from experimental groups were compared by student's t-test or ANOVA one way analysis. A value of p,0.05 was considered statistically significant.

The Proximal Region of SCN5a Promoter Contains Foxo1 Responsive Elements
Silico promoter analysis on a list of 354 of the most differentially expressed genes in the thymocyte lineage has revealed putative direct targets that are transcriptionally regulated by Foxo proteins in thymocytes, and that SCN5a transcription is upregulated by Foxos deletion [33]. These findings suggest that transcription of SCN5A may be directly regulated by Foxos in cardiomyocytes. Analysis of the SCN5a promoter region showed a highly conserved DNA sequence (59-CAAAACA-39) as IRE in humans and rodents such as mouse and rat. The mouse and human SCN5a promoter regions harbor two overlapping Foxo1 binding IREs, 59-CAAAACAAAACA-39 (Fig. 1A).

Foxo1 directly binds SCN5a promoter
To confirm the association of Foxo1 with the SCN5a promoter region including IREs, a ChIP assay was carried out in HL-1 cells and mouse ventricular myocytes, and it unveiled a clear enrichment of DNA amplified by PCR with two sets of primer pairs flanking Foxo1 binding IREs (59-CAAAACAAAACA-39) in the proximal SCN5a promoter region following immunoprecipitation with Foxo1 antibody, but not with IgG control (Fig. 1B). Another pair of primers was designed to amplify a region away from Foxo1 binding site, 59-CAAAACA-39. As showed in Figure  S1, there was no PCR product detected after chromatin immunoprecipitation by Foxo1 antibody.

Foxo1 Negatively Regulates SCN5a Promoter Activity
This conserved sequence is located from 21484 to 21473 in the mouse SCN5a promoter region. The proximal promoter region (22308 to +119 bp) was examined in luciferase reporter assays in HeLa cells with and without expression of Foxo1 ( Fig. 2A). Expression of Foxo1 significantly suppressed SCN5a promoter activity (p,0.01), and the suppression was Foxo1 dose dependent (Fig. 2B). Individual (CAAAAC to CCCAAC) or double mutations (CAAAACAAAACA to CCCAACCCAACA) in Foxo1 binding  sites significantly suppressed (p,0.01) Foxo1 responsiveness compared with wild type SCN5a, and double mutations of Foxo1 were significantly more effective (p,0.05) than mutation 1 on SCN5a promoter activity (Fig. 2C).
Foxo1 Negatively Regulates Na V 1.5 Expression in HL-cells HL-1 cells harbor robust, rapidly activated and inactivated Na + currents [39], indicating that Na V 1.5 is expressed in HL-1 cells. In order to test the hypothesis that Foxo1 regulates Na V 1.5 expression, HL-1 cells were used in the study. The HL-1 cells were infected for 48 hours by 10 or 50 multiplicity of infection (MOI) adenoviruses (Adv)-CMV-Foxo1-AA-GFP [40] and 50 MOI Adv-CMV-GFP. Western blot analysis of the protein extracts from these infected HL-1 cells was performed, and showed that Na V 1.5 protein was more significantly decreased (p,0.05 or p,0.01) in cells expressing constitutively nuclear Foxo-1-AA-GFP (AA: T24A/S319A) compared to cells expressing GFP ( Fig. 3A and C). Interestingly, endogenous Foxo1 was significantly increased by overexpression of Foxo1-AA-GFP (Fig. 3A). This finding is consistent with a previous report that Foxo1 can be positively regulated by Foxo1 binding its own promoter [41]. Accordingly, Na V 1.5 mRNA level determined by semi-quantitative RT-PCR analysis was also significantly decreased (p,0.01) in HL-1 cells expressing Foxo1-AA-GFP compared to HL-cells expressing GFP (Fig. 3B and D). This pair of primers (m1Na V 1.5) for RT-PCR application was designed to detect two splicing variants, Na V 1.5a and Na V 1.5b [42]. In order to determine if Foxo1 affects the expression of these two splicing variants, the second pair of primers (m2Na V 1.5) was designed to allow us to detect Na V 1.5, Na V 1.5a and Na V 1.5b. Unfortunately, there was only one band of PCR product at 1300 bp detected, which consistently showed the inhibition of Na V 1.5 expression by Foxo1-AA-GFP in HL-1 cells ( Figure S2). This has prevented us from addressing if Foxo1 also affects the expression of other splicing variant(s).  As shown above, robust Na V 1.5 expression was observed in the HL-1 cells (Fig. 3A and B). To test the hypothesis that a decrease of endogenous Foxo1 expression upregulates Na V 1.5 expression, Adv-FoxO1-RNAi was used to infect HL-1 cells to express RNAi and silence Foxo1 expression (Fig. 4A). The downregulation of Foxo1 induced a significant increase in both Na V 1.5 protein (p,0.05, at 100 MOI) (4A and C) and mRNA levels (p,0.01, at 100 MOI) in HL-1 cells (4B and D) when compared with those obtained from the cells infected with Adv-scramble-RNAi. This FoxO1-RNAi construct encodes a 19-bp oligonucleotide (59-CGCCCCAGGTGGTGGAGAC-39) that is complementary to the mouse Foxo1 mRNA sequence (10-29 nt). This sequence has two nucleotides different from that of human Foxo1 mRNA and one nucleotide at the most 59 different from that of TR4 mRNAencoded by Nrc2c gene. In order to determine the specificity of this RNAi, human HeLa cells expressing endogenous Foxo1 [43] and HL-1 cells were infected with Adv-Foxo-RNAi and Adv-scramble-RNAi, respectively. RT-PCR analysis showed that Foxo1 mRNA in HeLa cells and TR4 mRNA in HL-1 cells were not altered by expression of Adv-Foxo1-RNAi ( Figure S3 and S4), indicating that RNAi specifically targeted mouse Foxo1 mRNA.

Foxo1 Nuclear Localization Reduces Na + Currents in HL-1 cells
To test the hypothesis that activation of Foxo1 reducing Na V 1.5 expression decreases functional Na + currents, whole-cell voltageclamp Na + current recordings were performed on HL-1 cells infected with 50 MOI Adv-CMV-GFP and 50 MOI Adv-CMV-Foxo1-AA-GFP for 48 hours, respectively. Foxo1-AA-GFP was localized in the nuclei (Fig. 5B), while GFP was seen in both cytoplasm and nuclei (Fig. 5A). Recording traces of rapidly activating and inactivating Na + currents from GFP and Foxo1-AA-GFP-expressing cells, respectively, 36 hours after adenovirus infection, are illustrated in Figure 5C and D. Na + currents were significantly inhibited by expression of Foxo1-AA-GFP, and peak Na + current density was significantly lower (p,0.01) in cells expressing Foxo1-AA-GFP (19.665.6 pA/pF, at 225 mV) than in cells expressing GFP (67.969.6 pA/pF, at 225 mV). Fifty mM H 2 O 2 was then added to the culture medium. Fluorescent and confocal images showed that 2 hours of treatment with H 2 O 2 promoted Foxo1-GFP to localize in nuclei, but not GFP (Fig. 6).

H 2 O 2 Promoting Foxo1 Nuclear Localization Reduces Na V 1.5 Expression
H 2 O 2 inhibiting Na V 1.5 expression is through the Foxo1 signaling pathway: Increase of Foxo1 nuclear localization by H 2 O 2 is expected to reduce Na V 1.5 expression. In order to confirm this assumption, RNA and protein were extracted from HL-1 cells treated with 25 mM H 2 O 2 for 48 hours. Analyses of RT-PCR products and Western blot results showed that both Na V 1.5 protein and mRNA levels were significantly decreased (p,0.05)   . Foxo1 inhibits Na + channel activity. Foxo1-AA-GFP was localized in HL-1 cardiomyocyte nuclei (B) while GFP was seen in both nuclei and cytoplasm (A). Whole cell Na + currents were recorded during 80 ms depolarizing voltage steps to potentials between 260 and +60 mV from a holding potential of 2100 mV in HL-1 cells expressing GFP or Foxo1-AA-GFP for 36 hours. The typical whole cell recording traces showed that robust Na + currents were present in HL-1 cells expressing GFP (C) while remarkable reduction of Na + currents in these cells expressing Foxo1-AA-GFP (D). doi:10.1371/journal.pone.0032738.g005

Discussion
Foxo1 negatively regulates Na V 1.5 expression Foxo1 is known to function primarily as a transcription factor. The results presented here demonstrate that Foxo1 directly binds the IRE in the SCN5a promoter region and inhibits its promoter activity. This inhibition leads to downregulation of Na V 1.5 expression and a decrease of cardiac Na + channel activity. In the nucleus, Foxo proteins bind a conserved DNA binding sequence, 59-GTAAA(C/T)A-39 [30] or IRE, 59-CAAAA(C/T)A-39 [30][31][32] in the promoter regions of the targeted genes. Foxo recruitment to the promoter regions usually positively regulates expression of the target genes. For example, Foxo1 bound the consensus sequence in the Kir 6.1 gene promoter and increased the expression of KATP subunit Kir6.1 [25]. In the literature, there is one report in which the authors identified Foxo1 as a corepressor for androgen receptor (AR) [44]. Some corepressors possess intrinsic histone deacetylases (HDAC) activity and some repress gene expression through recruitment of HDAC proteins [45]. It has been shown that HDAC3, but not HDAC1 and HDAC2 associate with Foxo1, enhancing Foxo1-induced inhibition of AR function in LNCaP cells [46]. Whether Foxo1 recruits HADC3 to inhibit the SCN5a promoter activity remains undetermined.
H 2 O 2 suppressing Na V 1.5 expression requires Foxo1 ROS have been shown to acutely enhance late Na + currents, leading to action potential prolongation and early afterdepolarizations [47][48][49] via activating CamKII [8,7] and phosphorylating Na V 1.5. ROS have also been reported to acutely reduce fast transient outward K + current, leading to the prolongation of cardiac repolarization by phosphorylating Kv4.3 [50]. The chronic effects of H 2 O 2 on Na V 1.5 expression were also well studied [9]. Na V 1.5 expression was suppressed by H 2 O 2 activating NF-KappaB and partially inhibiting SCN5a promoter activity [9]. These findings indicate that other mechanisms are also involved in the regulation of Na V 1.5 expression at the transcriptional level. As reflected in our studies, H 2 O 2 reduced Na V 1.5 expression by promoting Foxo1 nuclear localization, and Foxo1 signaling is required for H 2 O 2 to affect Na V 1.5 expression. ROS modulate Foxo1 subcellular localization through different pathways. H 2 O 2 inhibits phosphorylation of Foxo1 at Ser-253 and The-24 by reducing Akt kinase activity in response to insulin [12]. Under oxidative stress, the activated JNK pathway decreases the activity of Akt in HIT cells, leading to the decrease of phosphorylation of Foxo1 and promoting its nuclear localization [51]. Oxidative stress also augments PRMT1-induced Foxo1 methylation, thereby leading to the inhibition of its phosphorylation by Akt and promotion of its nuclear localization [13]. Sirt1 has been reported to be activated by H 2 O 2 and the subsequently activated Sirt1 deacetylates and entraps Foxo1 in the nucleus, which is also independent of H 2 O 2 -induced dephosphorylation of Foxo1 [12]. It is likely that Sirt1, Akt, JNK and PRMT1 signaling pathways are involved in the regulation of Na V 1.5 expression under oxidative stress. Foxo4 has been shown to interact with NF-KappaB and subsequently inhibit its transcriptional activity [52]. Whether Foxo1 suppressing Na V 1.5 expression in cardiac myocytes is also partially through the NF-KappaB signaling pathway remains to be verified.

Novelty and Potential Significance
ROS including H 2 O 2 are significantly elevated in ischemic heart disease (IHD), [1][2][3][4] in which arrhythmias are frequently observed. A significant decrease in Na + channel activity is considered to be one of mechanisms contributing to the development of arrhythmias in IHD [5,6]. The mechanisms of transcriptional regulation of Na V 1.5 expression are largely uncertain. In this study, we demonstrated that Foxo1 directly targeted the SCN5a promoter and inhibited its transcriptional activity. Furthermore, we confirmed that H 2 O 2 promoted Foxo1 nuclear translocation. This Foxo1 relocation is required for H 2 O 2mediated Na V 1.5 downregulation. These findings indicate the important role of Foxo1 in cardiac ion channel regulation and in the development of arrhythmias.

Supporting Information
Figure S1 Foxo1 does not bind the SCN5a promoter region lacking the insulin responsive element. A pair of primers was designed to amplify a region far away from Foxo1 binding site, 59-CAAAACA-39. There was no PCR product detected after chromatin immunoprecipitation (ChIP) by Foxo1 antibody (Lane 3). Lane 1 and 2 were input DNA and control IgG ChIP, respectively. (TIF) Figure S2 One isoform of Na V 1.5 is detected in HL-1 cells. RT-PCR using a pair of m2Na V 1.5 primers showed that there was only one DNA band representative of Na V 1.5 mRNA after 40 cycles. Overexpression of 50 MOI Foxo1-AA-GFP for 36 hours decreased Na V 1.5 mRNA level compared with that in cells expressing 50 MOI GFP and the amount of sample loading was equal as determined by GAPDH RT-PCR products. (TIF) Figure S3 RNAi targeting mouse Foxo1 mRNA does not affect human Foxo1 expression. RT-PCR using a pair of hFoxo1 primers showed that human Foxo1 mRNA expression was not altered in the HeLa cells infected with 50 MOI Adv-Foxo1-RNAi (n = 3) in comparison with the cells infected 50 MOI Advscramble-RNAi (n = 3) (A and B). (TIF) Figure S4 RNAi targeting mouse Foxo1 mRNA does not affect TR4 mRNA expression. RT-PCR using a pair of TR4 primers showed that TR4 mRNA expression was not altered in the HL-1 cells infected with Adv-Foxo1-RNAi (n = 3) in comparison with the cells infected Adv-scramble-RNAi (n = 3) (A and B). (TIF)