Forkhead Transcription Factors Coordinate Expression of Myocardial KATP Channel Subunits and Energy Metabolism
Coordinate adaptation of myocyte metabolism and function is fundamental to survival of the stressed heart, but the mechanisms for this coordination remain unclear. Bioinformatics led us to discover that Foxs are key transcription factors involved. We performed experiments on the mouse atrial cell line HL-1, neonate rat heart myocytes, and an adult rat model of myocardial infarction. In electrophoretic mobility-shift assays, FoxO1 binds to the FoxO concensus site of the KATP channel subunit KIR6.1 promoter. In primary atrial culture, targeting FoxO1 and FoxO3 with siRNA specifically reduces mRNA expression of FoxO1 and -O3 and KIR6.1. Western blots, confocal immunofluorescence, and quantitative RT-PCR was applied for measuring expression of 10 Fox, 6 KATP channel subunits, and 12 metabolic genes. FoxF2, -O1, and -O3 strongly associate with expression of KATP channel subunits (in particular, KIR6.1, SUR1A and SUR2B) in different heart tissues and in the periinfarct zone of the left ventricle. Patch-clamp recordings demonstrate that molecular plasticity of these channels is matched by pharmacological plasticity and increased sensitivity to a metabolic challenge mimicked by the protonophore CCCP. A balance of FoxF2 and FoxO also regulates expression of at least 9 metabolic genes involved in setting the balance of glycolysis and β-oxidation. Bioinformatics shows that the transcriptional mechanisms are highly conserved among chicken, mouse, rat, and human, and Fox are intimately linked to other metabolic sensors. Thus, FoxF2 and -O are key transcription factors coordinating expression of KATP channels and energy metabolism.
Metabolic adaptation is vitally important for cellular function, stress adaptation, and survival. This is particularly true in metabolically highly active tissues. In heart,1 brain,2 kidney,3 adipose tissue,4 muscle,5 and blood vessels,6 the ATP-dependent potassium channels (KATP channels) emerge as the most important sensors of cellular energy status.7 During metabolic stress, these heterooctameric channels open in response to an increased cytoplasmic ADP/ATP ratio, thereby hyperpolarizing the cell and reducing the calcium influx and metabolic demand.1 KATP channels protect cardiac myocytes in pathophysiological situations,8 as well as during vigorous exercise. For example, mice with knockout of the pore-forming KATP channel subunit KIR6.2, or mice overexpressing dominant negative KIR6.1, tolerate only half of the workload compared with wild-type mice, and half of them die.9,10
KATP channels are, however, amazingly diverse, varying in their subunit composition from one tissue to the next. The potassium channel pore consists of 4 KIR6.2 subunits in left ventricle,11 pancreatic β-cells,12 and vascular endothelium13 and mainly of 4 KIR6.1 subunits in vascular smooth muscle.14 The brain2 and skeletal muscle15 express both KIR6.1 and KIR6.2. The same cardiac tissue may change its subunit composition: subunit Kir6.1 is increasingly expressed in left ventricle following ischemia,16 exercise,17 or treatment with KATP channel agonist.18 Moreover, differences in subunit composition of cardiac KATP channels arise during development, aging, and between male and female.19–21 The varying expression of the regulatory subunit sulfonylurea receptor (SUR) adds additional complexity. SUR2A is expressed in left ventricle,22 SUR1A in pancreatic β-cells, and SUR2B in vascular smooth muscle.23
The puzzling molecular diversity of these metabolic sensors raises 2 unresolved questions. What mechanisms cause the varying expression of the KATP channel subunits? Is altered composition of KATP channels coordinated with altered metabolic control? We previously found that atria and ventricles strikingly differ in KATP channel properties and sensitivity to ADP24,25 and sought to determine whether corresponding differences existed in transcriptional mechanisms. Accordingly, our bioinformatics approach reveals that Forkhead transcription factors (Foxs) would be prime candidates mediating KATP channel expression (Figure 1 and Tables 1 and 2⇓). Foxs have been known since 1990. Their specific DNA-binding domain is highly conserved through animal evolution. They are emerging as key players in developmental and metabolic processes and could act as transactivators or transrepressors.26 Their powerful role is attested by the fact that some Foxs, such as FoxA, can act independently of acetylation or compaction of the nucleosome. In pancreatic β-cells, FoxA2 causes expression of a cluster of genes including SUR1-KIR6.2, GLUT2, glucokinase, hexokinases, and l-pyruvate kinase.27 The number of Fox members seems to correlate with anatomical and functional complexity: there are 4 Fox genes in Saccharomyces, 15 in Caenorhabditis, 20 in Drosophila, and 39 in Homo sapiens.26 Potential binding sites for Foxs also exist on many genes encoding glucose and fatty acid transporters and enzymes of the glycolytic and β-oxidation pathway (Table 3). Thus Foxs possibly coordinate the expression of KATP channels and enzymes and transporters of metabolic pathways. The experiments described below support this hypothesis.
Materials and Methods
Hearts from 2- to 3-day-old rats were carefully microdissected to separately collect right atrial appendages, left atrial appendages, left ventricles, and right ventricles. The tissues were incubated for 4 hours at 4°C in trypsin (2.5%) and dissociated for 15 minutes at 37°C. After preplating at a density of 130 000 cells/cm2 on plastic to remove a majority of fibroblasts, the suspended cells were plated and cultured for 24 hours on plastic dishes for the extractions of mRNA and protein. Cells were also plated at low density (10 000 to 50 000 cells/dish) on fibronectin–gelatin coated 22-mm glass slides and cultured pending immunocytochemistry or patch-clamp recordings, as described previously.28,29 Medium was DMEM-F12 with 10% FBS, insulin–selenium–transferrin, penicillin–streptomycin, and ascorbic acid. To prevent fibroblast proliferation, 20 μmol/L cytosine arabinonucleoside was added to the culture medium. Contamination with vascular smooth muscle cells or myofibroblasts was 9.1±2.2% after 48 hours (mean±SEM, N=10), as ascertained by immunofluorescence for α-smooth muscle actin (antibody was provided by Dr M.-L. Piallat, Centre Médical Universitaire, Geneva, Switzerland).
The HL-1 mouse atrial cell line30 was a generous gift from Dr William C. Claycomb (LSU Health Sciences Center, New Orleans, La). HL-1 cells were cultured under a 5% CO2 atmosphere in Claycomb medium (JRL Bioscience) supplemented with 10% FBS, 4 mmol/L l-glutamine, 0.1 mmol/L norepinephrine, 100 U/mL penicillin, and 100 μg/mL streptomycin. The medium was replaced every 24 to 48 hours. Cells were grown in T75 culture flasks or on 22-mm glass coverslips that were precoated overnight with Dulbecco’s PBS containing 0.012 mg/mL fibronectin and 0.2 mg/mL gelatin.
This investigation conforms to the NIH Guide for the Care and Use of Laboratory Animals. Myocardial infarction was induced in 200- to 225-g male rats by ligature of the left coronary artery as described.31,32 Control rats underwent a sham operation. One, 7, 28, and 140 days after coronary occlusion, the heart was excised under deep pentobarbital anesthesia (150 mg/kg IP). The infarct border zone, defined as the myocardium 2 mm around the infarct scar, was carefully dissected out. Cardiomyocytes were isolated by incubation in collagenase solution, followed by stepwise recovery to physiological calcium levels.
RNA Extraction, Reverse Transcription, and Real-Time PCR
Total RNA was extracted by disrupting cells in TRIzol reagent (Invitrogen) and prepared according to recommended procedures. The quality of RNA was controlled in agarose gels by the 18S/28S ratio. RNA concentration was measured by spectrophotometry at 260 nm, and their quality was controlled by the 280/260 nm emission ratio. RNA was subjected to DNase I to remove all genomic DNA contamination with DNA-free (Ambion) as indicated by the manufacturer. Reverse transcription was performed on 1 μg of total RNA with Thermoscript (Invitrogen), random primers, and dNTP as indicated by the manufacturer. Real-time PCR was performed on cDNA with SYBR Premix ExTaq Takara in an I-Cycler thermocycler (Bio-Rad). Oligonucleotides were designed with the software Primer express 2. Specific TaqMan probes were synthesized by Applied Biosystems and primers by Microsynth (for primers and probes, see Table 4⇓). The standard curve method was used for relative quantification. The mRNA values were normalized by the corresponding 18S rRNA and cyclophilin mRNA. The different normalization procedures yielded similar results.
Electrophoretic Mobility-Shift Assays
Nuclear extracts from cell cultures were prepared according to the method of Schreiber et al.33 Hela cells were used as positive control of the method, using probes for SP-1. Transfected HL-1 cells and nontransfected primary cultured atrial myocytes were used for the tests. Positive controls for Fox binding were made with HL-1 cells transfected with FoxO1 plasmid. Oligonucleotides were obtained from Microsynth: ATTCGATCGGGGCGGGGCGAG for the transcription factor SP-1 as a positive control; TGATGAGTGTTTGTTTATGAG (FoxO probe; Figure 4A) and GATGAGTGTTTATTATATGAG (FoxO1 probe; Figure IC and ID in the online data supplement) for the consensus sequence (mouse, rat, human conserved) found in the KIR6.1 promoter. Oligonucleotides were labeled by the T4 polynucleotide kinase with [γ-32P]ATP and purified on spin columns (NucleoSpin extract II [MN]). The labeled probes (80 000 to 100 000 cpm) were incubated with, or as a control without, the nuclear extracts in gel shift–binding buffer (20% glycerol, 5 mmol/L MgCl2, 2.5 mmol/L EDTA, 2.5 mmol/L dithiothreitol, 250 mmol/L NaCl, 50 mmol/L Tris-HCl (pH7.5), 0.25 mg/mL poly(dI-dC). Cold probe was coincubated for displacing the radioactive probe. Electrophoresis of DNA–protein complexes was performed on 4% polyacrylamide gels (60:1 acrylamide:bisacryl-amide). Gels were dried and placed for at least 5 days against an autoradiographic film with an amplification screen at −80°C.
Cell cultures were disrupted with lysis buffer (25 mmol/L Tris, 150 mmol/L NaCl, 5 mmol/L EDTA, 1% [vol/vol] Triton X-100, pH 7.5, supplemented with a cocktail of protease inhibitors [Roche]). Fifty micrograms of protein were loaded on a polyacrylamide–sodium dodecyl sulfate gel that was blotted on a 0.45-mm nitrocellulose membrane BA85 (Schleicher et Schuell), as described previously.33a Loading controls were performed by red Ponceau staining. Hybridizations were revealed with chemiluminescent substrate according to the protocol of the manufacturer (Roche). Quantification of the detected proteins was performed by scanning and with Metamorph software (Universal Imaging). Primary antibodies were Goat anti-Kir6.1 (sc-11224, Santa Cruz Biotechnology), rabbit anti-Kir6.2 (APC-020, Alomone), rabbit anti-SUR (sc-5791, Santa Cruz Biotechnology), rabbit anti-SUR2B (sc-5793 Santa Cruz Biotechnology), rabbit anti-FoxF2 (ab23306, Abcam), goat anti-FoxO3 (ab17026 Abcam), rabbit anti–green fluorescent protein (GFP) (Synaptic Systems).
FoxF2 expression plasmids were a generous gift from Dr P. Carlsson, Göteborg University (Göteborg, Sweden).34 Wild-type FoxO1-GFP expression plasmids were a generous gift from Dr T.G. Unterman (McGill University, Montreal, Canada).35 FoxO3a expression plasmids were ordered at ADGENE and are the original construct from Dr M.E. Greenberg (Harvard Medical School, Boston, Mass).36
Small Interfering RNA
Small interfering (si)RNAs were obtained from Invitrogen to knock down mouse FoxO1 and -O3 and were successfully tested in HL-1 mouse atrial myocytes. Because the design was optimized to also knock down rat Fox, 1 of the siRNA for FoxO1, and 2 of the siRNAs for FoxO3 also worked in rat. siRNAs with scrambled sequence served as control. A fluorescent siRNA was used as a control for successful electroporation. For each culture well, 1 electroporation (see below) was performed immediately after cell dissociation on 2 million preplated cells. Up to 6 wells were pooled for RNA extraction from the cultures 24 hours later. Each pooled sample served as 1 single data point, and 5 to 7 data points per gene were averaged for the statistical analysis.
Cell Transfection and Electroporation
Lipofections of HL-1 cells were performed with Lipofectamine 2000 (Invitrogen) as recommended by the manufacturer. Cells were lysed 48 hours posttransfection. For single-cell studies, primary cultured myocytes were electroporated with a Bio-Rad Genpulse electroporator as described.28 For nucleoporation of primary cultured atrial myocytes with siRNAs, protocol no. 09 of Amaxa AG (Köln, Germany) was applied, as recommended by the manufacturer.
Immunocytochemistry, Confocal Microscopy, and Quantitative Imaging
Myocytes were fixed in 2% paraformaldehyde and stained as described,28 using antibodies listed previously. Image stacks of immunostained cultures were acquired on a confocal microscope at a fixed laser setting and analyzed.28 For each cell, the mean background-subtracted fluorescence of regions of interest (nucleus, cytoplasm) was determined with Metamorph software at 3 z-positions, normalized by exposure time, and averaged; means of averages±SEM were obtained for groups of 10 to 16 cells. For example, red fluorescence of KIR6.1/Texas Red was plotted as function of cytoplasmic or nuclear green fluorescence of FoxO1–enhanced GFP (EGFP).
Promoter analysis was performed online at the Lawrence Livermore Institute Laboratory (http://www.dcode.org) for the following genes: SUR1-KIR6.2, SUR2-KIR6.1, MLC2a, MLC2v, proANP, proBNP, and the metabolic genes listed in Tables 1 and 2⇑. The analysis of the promoter across species was performed for chicken, mouse, rat, and human with the ECR Browser software (http://ecrbrowser.dcode. org). The alignment between pairs of promoter regions was performed with blastz software and designed with zpicture (http://zpicture.dcode.org). This alignment was submitted to Rvista2 to identify the putative transcription factor binding sites at http://rvista.dcode.org. The putative promoter and transcription factor binding regions were confirmed with Genomatix ElDorado and Gene2Promoter software.
Whole cell patch-clamp recordings of the KATP current were obtained as described previously.24,25,29 From a holding potential of −40 mV, voltage ramps were imposed every 30 seconds from −80 to +90 mV over a 10-second period. This resulted in quasi–steady-state current–voltage curves. Membrane potential was measured in current-clamp mode at 0 pA at the end of each ramp. The pipette solution contained (in mmol/L) 120 KCl, 1.3 CaCl2, 1.3 MgCl2, 10 Hepes, 10 glucose, 10 BAPTA, plus 1 mmol/L K-ATP and 10 μmol/L K-ADP. The pH was adjusted to 7.3, and osmolality to 290 milliosmol/kg. The bath solution contained (in mmol/L): 5 KCl, 1 CaCl2, 1 MgCl2, 118 NaCl, 10 Hepes, and 10 glucose. The pH was adjusted to 7.4, and osmolality was adjusted to 290 milliosmol/kg with sucrose. Drugs were administrated to 23 rat cardiomyocytes in the perfusion system in the sequence control, diazoxide (100 μmol/L), pinacidil P-1075 (100 μmol/L), glibenclamide (1 μmol/L), with 2 to 4 minute washes between drug applications. Metabolic sensitivity was tested in a separate series of 24 rat cardiomyocytes by application of low concentrations (20 nmol/L) of the protonophore CCCP to mimic metabolic stimulation.25
The statistical analysis was performed on Sigmastat 2 SPSS. The mRNA and protein levels were compared by ANOVA with a post hoc Newman–Keuls test. Time varying expressions of Foxs in cardiac infarction studies were compared with a Dunnett test. The correlation analysis was performed with a linear regression and a covariance test, and confirmed by Spearman correlation. Multiple regression analysis was performed with SPSS software.
Abundance of Fox-Binding Sites on SUR-KIR6 and Metabolic Genes
Alignments of chicken, mouse, rat, and human genes coding for KATP channel subunits and metabolic enzymes allowed us to identify the most highly conserved sequences in the 5′ upstream region and the presumed transcription factor binding sites (Figure 1; detailed charts are available on request). In sequences 5′ upstream of the Kir6.1 coding region, Rvista2.0 finds a high redundancy of binding sites for FoxO1, -O3, -O4, -F2 (FREAC2), -C1 (FREAC3), -D1 (FREAC4), -L1 (FREAC7), and a common Forkhead consensus site. Immediately upstream of the SUR2 coding region, Rvista2.0 finds almost the same sites: FoxO1, -O3, -F2, -D1, -D3, -J2, -P1, and a common Forkhead consensus site. For the Kir6.2 gene, Rvista finds only 1 putative site for FoxC1 (FREAC3) and high redundancy for MEF2, OCT, SREBP1a, SP1, AP2, and NF-κB.
Presumed promoter sequences of KATP channel subunits were also aligned with presumed promoter sequences of proANP, proBNP, MLC2a, and MLC2v (Tables 1 and 2⇑; detailed charts available on request). There are 15 highly conserved sequences 5′ upstream of the Kir6.1 gene (more than 90% sequence identity between species) but only 2 sequences 5′ upstream of the SUR2 gene. Alignments of the KIR6.1 promoter with putative atrial markers (MLC2a and proANP promoters) yield striking clusters of binding sites for FoxA1, -A2, -C1, -D3, -F1, -F2, -I1, -J1, -J2, and -L1 (Tables 1 and 2⇑). Alignments with the MLC2v and proBNP promoter regions yield negative results, except for FoxP1.
A similar analysis on metabolic genes yields clusters of Fox binding sites on the CD36, LPL, FABP3, MCAD, ACCβ, and MCD promoters for the β-oxidation pathway and on the GLUT4 and PKM promoters for the glycolytic pathway (Table 3). Interestingly, a Foxs cluster is also observed for SREBP1a, as are binding sites for SREBP1a, HIF1α, PPAR, and Nkx2.5. Thus in silico, Foxs emerge from a library of 430 transcription factors as the most prevalent transcriptional regulators of KATP channel subunits and metabolic pathways.
FoxO1, -O3, and -F2 Cause Increased Gene Expression of KIR6, SUR, Metabolic Transporters, and Enzymes in Atrial Myocytes
FoxO1, -O3, and -F2 were selected from a group of 10 Foxs from ongoing experiments on rat hearts. The transfection of HL-1 cells with a FoxO3 plasmid36 induces a 15-fold increase in FoxO3 mRNA, a 3-fold increase in KIR6.1 mRNA, a tight positive correlation between FoxO3 and KIR6.1 protein in quantitative double-immunocytochemistry, and a 5- to 26-fold increase of KIR6.1 protein in Western blots (Figures 2 and 3⇓). Moreover, FoxO3 induces a 2-fold increase in KIR6.2 and a 4-fold increase in SUR2A mRNA (Figure 2). Wild-type FoxO1-EGFP35 induces green fluorescence and increased KIR6.1 expression (supplemental Figure IA and IB; R=0.66, P<0.01, n=22 cells), and a 6-fold increase in KIR6.1 protein (Figure 3). Interestingly, transfection with FoxO1-EGFP (or FoxO3) represses FoxF2 protein, whereas human FoxF2 plasmid34 increases FoxO1/O3 mRNA and protein 10-fold (Figure 3B). FoxF2 also increases SUR2A, KIR6.2, and KIR6.1 mRNA (Figure 2). In primary rat atrial myocytes, electroporation with FoxO1-EGFP plasmid induces green fluorescence and increased expression of immunoreactive KIR6.1 (Figure 3G); electroporation with FoxF2 plasmid increases nuclear FoxF2 in inverse correlation with nuclear FoxO1-EGFP (R=−0.51, P<0.05, N=16), supporting the results from HL-1 cells.
Foxs potently modulate the expression of genes coding for free fatty acid and glucose transport and metabolic enzymes (Figure 2). Many of these findings are new. FoxF2 downregulates, whereas FoxO1 induces phospho-fructokinase 2, in HL-1 cells, possibly via downregulation and induction of SREBP1a, respectively.37–39 Overexpression of FoxF2 downregulates gene expression throughout the glycolytic pathway. FoxO1 and -O3 induce increased expression of malonyl-CoA decarboxylase (MCD), which inactivates malonyl CoA, an inhibitor of carnitine palmitoyltransferase-1 (CPT-1), and FoxO3 increases expression of medium chain acyl CoA dehydrogenase (MCAD). The metabolic sensor AMPKα1 is also upregulated by FoxO1 and -O3. Additional findings confirm previous studies on other cell types40 and LPL41; gluconeogenesis, glycolysis, and lipid gene expression.39 Putative Fox binding sites also exist on peroxisome proliferator-activated receptors (PPARs), which positively impact on fatty acid oxidation, as confirmed by experiments on HL-1 cells. The results thus suggest cause-and-effect between Foxs and expression of KATP channels and metabolic genes.
Causal Relationships Between FoxO and Channel Subunit Expression Are Demonstrated by Electrophoretic Mobility-Shift Assays and siRNAs
Nuclear extracts from mouse atrial HL-1 myocytes overexpressing FoxO1 slow down the gel migration of a P32-labeled oligonucleotide with FoxO consensus sequence that is present in the KIR6.1 gene promoter. Cold probe erases this shift (Figure 4A, top). Nuclear extracts from nontransfected primary rat cardiomyocyte cultures also cause a shift of the radioactive FoxO probe, and the cold probe decreases the intensity of the shifted band by >90% (supplement Figure IC and ID). Cause-and-effect is further indicated by a 76% to 83% decrease in expression of KIR6.1 mRNA, 24 hours following electroporation of primary cultured rat atrial myocytes with anti-FoxO siRNAs (Figure 4B). Remarkably, anti-FoxO1 siRNA affects primarily FoxO1 and less so FoxO3 mRNA, and vice versa, but both types of siRNAs inhibit KIR6.1 expression relative to the control with scrambled sequence.
Coordination of Tissue-Specific Expression of Fox, Channel Subunits, and Metabolic Genes in Neonate Rat Cardiomyocytes
Atria and ventricles differ in their development and mechanical stress; right and left heart are exposed to different oxygen tension. Would they differ in Fox, channel subunits, and metabolic genes? This is indeed the case. FoxA2, -C2, and -F2 display the highest mRNA levels in the left ventricle, whereas FoxO1, -O3, -O4, and -P1 are highly expressed in the atria (Figure 5A). FoxM1 is expressed at the same level in all cardiac compartments (supplemental Figure II). There is a right/left heart difference in expression for FoxJ2. At the protein level, FoxO3 protein is expressed more strongly in right atrium and FoxF2 in left ventricle. This is indicated by Western blots (Figure 6) and quantitative immunocytochemistry (supplemental Figure IIIA). FoxO3 is mainly localized in the nucleus. FoxF2 resides in both the nucleus and perinuclear zone (Figure 6A2 and 6B2).
Significant associations exist between Fox and KATP channel subunit mRNA: both positive (in black: FoxJ2, -O1, -O3) and negative (in gray: FoxD3, -F2) (Figures 5B, 6A3, and 6⇑B3; details in Table 5). Atria express 4- to 12-fold higher levels of KIR6.1 and SUR1A mRNA than ventricles, and right heart expresses up to 6-fold higher levels of SUR2B mRNA than left heart (Figure 7A). SUR2A, KIR6.2,l and SUR1B mRNAs are evenly distributed throughout the cardiac chambers. Significant associations also exist between Fox and metabolic enzyme mRNAs (Table 5). In general, ventricles express higher mRNA levels for all enzymes and transporters studied (Figure 8); many of these findings are new, and others extend previous results.42,43
Protein levels generally conform to the mRNA measurements (Figure 7B). Kir6.1 is significantly (2-fold) higher in atrial than ventricular myocytes, whereas KIR6.2 is evenly distributed throughout. Immunocytochemistry shows the presence of KIR6.1 protein in myocytes, displaying the typical striated pattern in left ventricular myocytes (Figure 7C), a patchy pattern in atrial myocytes, and a mixed pattern in right ventricle. These most probably reflect the different amounts of contractile proteins present in neonatal myocytes. SUR2B expression is highest in atria and very low in left ventricle (supplemental Figure IIIB).
High Expression of Foxs Following Myocardial Infarction
Results so far indicate a tissue-specific, Fox-related expression of KATP channel subunit and metabolic genes, possibly caused by the varying metabolic challenges. Would ventricular tissue, when exposed to stress, acquire new Fox-related characteristics? Because FoxO1 and -F2 knockout mice are not viable, we tested this hypothesis in a rat model of myocardial infarction,31 where KIR6.1 and SUR mRNAs are highly expressed in the infarct border zone.44 This zone is exposed to low Po2 and high mechanical stress.45 We tested in these same rats the time-varying patterns of Fox expression. Myocardial infarction elicits waves of Fox expression, starting with FoxC2 and -O1 at 7 days, continuing with FoxO3 and -J2 at 56 days, and FoxF2, -O3, and -J2 at 20 weeks (open circles, Figure 9A). Overexpression of genes in the periinfarcted zone reaches mRNA copy numbers found in the Fox-transfected HL-1 cell line. No significant changes occur in FoxA2, -O4, -D3, -P1 (Figure 9A) and -M1 (supplemental Figure IVA). In shams (closed circles), all Fox mRNA levels remain stable over the 20 week period.
Changes in Foxs parallel changes in KATP channel subunit mRNA. The most striking results are the strong positive associations between FoxF2 and KIR6.1, SUR1A, SUR2A, and SUR2B (Figure 9B and Table 6), and associations with FoxO1, -O3, -J2, and -C2. These represent almost the same set of transcription factors revealed in studies on neonate rat myocytes. Note that at these high levels of FoxF2 expression, there is a high expression of KIR6.1, as in HL-1 cells.
Other potassium channels, such as IK1 (KIR2.1, KIR2.3),46 Ito (Kv1.4),47 IKs (KvLQT1),48 and IKAch (KIR3.1),49 also display a regional expression in the heart. However, Rvista bioinformatics analysis of their 5′ promoter/enhancer regions fails to suggest a clear implication of Foxs. Indeed, no association was found between Foxs and KIR2.1 or KvLQT1 mRNA after myocardial infarction (supplemental Figure IVB and IVC).
Functional Plasticity Matches Molecular Plasticity of KATP Channels
A remarkable molecular plasticity thus exists for Foxs and channel subunits, indicating a pronounced expression of subunits KIR6.1, SUR1A, and SUR2B in regions (right atrium, left ventricle infarct border zone) exposed to low Po2. Atria are already known to display increased diazoxide sensitivity in correlation with increased metabolic sensitivity of the KATP channels,25 but right/left differences had not yet been examined. Right atrial myocytes from neonatal rats readily responded to 100 μmol/L diazoxide by generating within 3 to 4 minutes an inwardly rectifying potassium current of 250 to 550 pA (eg, Figure 10A1). Subsequent 100 μmol/L pinacidil (P-1075) only slightly increased the current by 200 pA. The potassium current was totally abolished by 1 μmol/L glibenclamide, as expected. Diazoxide (100 μmol/L) hardly induced any current in left ventricular myocytes (eg, Figure 10A2), whereas pinacidil induced a current comprised between 500 to 1000 pA (positive control). In Figure 10A4, the current densities clearly indicate a high sensitivity to diazoxide in right and left atrium and right ventricle, and a low sensitivity in left ventricle. In contradistinction, the current densities in response to pinacidil were statistically identical in all tissues (Figure 10A5). The myocyte KATP current in the highly stressed infarct border zone of the left ventricle is also highly sensitive to diazoxide.44 To better characterize the functional implications, another series of experiments was performed on 24 cardiomyocytes by mimicking a metabolic challenge with a very low concentration (20 nmol/L) of CCCP (Figure 10B). Brisk responses were obtained in right and left atrial and right ventricular myocytes, and a very small response in left ventricular myocytes. Thus normal left ventricular myocytes stand out by their minimal responses to diazoxide and CCCP. Compared with left ventricular myocytes, right ventricular myocytes express as much mRNA for creatine kinase, adenylate kinase, and pyruvate kinase (Figure 8); thus increased ATP-ADP buffering in left ventricular myocytes seems not to account for their lack of response to CCCP.
The capability of adapting cellular function and energy metabolism to varying physiological and pathological conditions is vitally important in animal cells. In the present work, we show that the Fox family may play a central role in expression of both molecular sensors of energy status and key regulatory genes of energy metabolism. First, in atrial cells, the expression of FoxO1, -O3, and -F2 cause increased expression of KATP channel subunits (the quintessential metabolic sensors7) and selective up- and downregulation of specific metabolic genes. A causal relationship between FoxO and KIR6.1 expression is demonstrated by electrophoretic mobility-shift assay and by experiments with siRNAs. Second, FoxO1, -O3, -F2, and -J2 are distributed unevenly within different cardiac chambers of the neonatal rat, in association with channel subunits KIR6.1, SUR1A, and SUR2B and 9 metabolic genes.42,43,50 Third, the periinfarcted zone of the rat left ventricle reveals an impressive plasticity of FoxO1, -O3, -F2, and -J2, associated with the increased expression of KATP channel subunits (KIR6.1 and all SUR) and with the known remodeling of metabolic pathways.31 These findings support the hypothesis that Foxs mediate the tissue-specific gene expression in the face of different mechanical and hypoxic challenges.
Although our study demonstrates a Fox-mediated expression of KATP channel subunits, it does not resolve their quantitative subcellular distribution in mitochondria, sarcoplasmic reticulum, and plasma membrane. In confocal microscopy, immunofluorescence signals from subunits that are endogenously expressed in mitochondria and reticulum overwhelm signals emitted from the adjacent plasma membrane. Thus definite proof is still lacking that sensitivity of atrial and right ventricular myocytes to diazoxide or CCCP is explained solely by increased plasmalemmal KIR6.1, SUR1A, or SUR2B.
Would plasticity in KATP channel expression and composition be advantageous to the cell? The total absence of cardiac KIR6.2 expression or inactivation by dominant negative KIR6.1 is clearly detrimental to the tolerance of intense exercise in mice.9,10 Thus increased expression and function of KATP channels carry vital benefits. The increased expression of KIR6.1 (a channel subunit of lower conductance than KIR6.251) and the expression of SUR1A and SUR2B may serve to limit the maximal KATP current while increasing the sensitivity to ADP52 and to a metabolic challenge. One may anticipate that following myocardial infarction, the periinfarcted zone will readily adapt to hypoxia. The risks of reentry arrhythmia are minimized, because the low conductance KIR6.1 subunit prevents an excessive shortening of the action potential.8,25,53 Interestingly, recent studies unrelated to KATP channels have shown that myocardial infarction,54 myocardial reperfusion injury,55 heart failure,56,57 or myocyte hypertrophy58 induce an overexpression of FoxO. Other studies unrelated to FoxO have shown increased expression of KIR6.1 following ischemia16 and exercise.17 Moreover, recent work suggests that FoxO and KATP channels reduce cardiac ageing and electrical instability in drosophila.59 Our results thus link tissue stress and expression of Fox and KATP channels into a coherent framework.
Other sensors of energy status impact on the expression and function of Foxs. AMP kinase (AMPK)60–62 may signal metabolic stress (AMP/ATP ratio) to Fox.63–65 We show that FoxO1 and -O3 strongly increase the expression of AMPKα1 in atrial cells; in hepatocytes, AMPK stimulation directs FoxO1 to the proteolytic degradation pathway,60 possibly in a negative-feedback arrangement. Sirtuins66 are activated by the NAD+/NADH ratio and trap FoxO1 in the nucleus of hepatocytes.67 The low-oxygen sensor HIF1α68,69 is a potential transcription factor for FoxC2, -F2, -O1, and -O3 (supplemental Figure V). Homologs of SREBP1a act as oxygen sensors in fission yeast70; SREBP1a is a potential transcription factor for expression of the SUR2-KIR6.1 and KIR6.2 genes (Figure 1). Although not exhaustive, this list of energy and low-oxygen sensors demonstrates their potential, intimate relationship with Foxs.
Based on present experiments and previous studies,27,71–73 we consider that each cardiac chamber is chronically exposed to a different workload, metabolic demand, and metabolic supply. This leads through as yet unidentified signal transducers to different expression profiles of Foxs and thus of KATP channel subunits and key enzymes for β-oxidation or glycolysis. In response to an increased workload, each cardiac chamber adjusts its local excitability, force of contraction, and mobilization of metabolic pathways, thus optimizing the overall adaptation of the heart to stress. In pathological situations, for example, after myocardial infarction, the chronic stretch45 or possibly recurrent ischemia in the infarct border zone remodels the left ventricular wall, changes the KATP channel profile, and modifies the expression level of key enzymes of β-oxidation and glycolysis. Controversial results regarding the switch from fatty acid oxidation to glycolysis32,74–77 may result from a variable interplay between FoxO3, -O1, -F2, -C2 and PPAR expression.
In conclusion, FoxF2, -O1, -O3, and most likely -C2 and -J2 are intimately involved in both metabolic sensitivity of KATP channels and transcriptional control of energy metabolism. They direct a tissue- and stress-dependent expression of metabolically sensitive potassium channels and the mobilization of additional or alternative energy supplies. The Fox-dependent coordination of metabolic responses appears to be of vital importance in metabolically highly active tissues. Identification of the exact signaling pathways from stress to Foxs remains a challenging task of the future.
We thank Irene Papageorgiou for excellent assistance in studies on myocardial infarction and Mauro Serafin for advice in patch clamping.
Sources of Funding
Supported by the Swiss National Science Foundation, Swiss University Conference project “Heart remodelling in Health and Disease,” Swiss Heart Foundation, Novartis Foundation, Société Académique de Genève, and the Gustave and Simone Prévot Foundation.
Original received June 12, 2007; resubmission received October 25, 2007; revised resubmission received December 5, 2007; accepted January 3, 2008.
Thabet M, Miki T, Seino S, Renaud JM. Treadmill running causes significant fiber damage in skeletal muscle of KATP channel-deficient mice. Physiol Genomics. 2005; 22: 204–212.
Tong X, Porter LM, Liu G, Dhar-Chowdhury P, Srivastava S, Pountney DJ, Yoshida H, Artman M, Fishman GI, Yu C, Iyer R, Morley GE, Gutstein DE, Coetzee WA. Consequences of cardiac myocyte-specific ablation of KATP channels in transgenic mice expressing dominant negative Kir6 subunits. Am J Physiol Heart Circ Physiol. 2006; 291: H543–H551.
Zingman LV, Hodgson DM, Bast PH, Kane GC, Perez-Terzic C, Gumina RJ, Pucar D, Bienengraeber M, Dzeja PP, Miki T, Seino S, Alekseev AE, Terzic A. Kir6.2 is required for adaptation to stress. Proc Natl Acad Sci U S A. 2002; 99: 13278–13283.
Babenko AP, Gonzalez G, Aguilar-Bryan L, Bryan J. Reconstituted human cardiac KATP channels: functional identity with the native channels from the sarcolemma of human ventricular cells. Circ Res. 1998; 83: 1132–1143.
Chatterjee S, Al-Mehdi AB, Levitan I, Stevens T, Fisher AB. Shear stress increases expression of a KATP channel in rat and bovine pulmonary vascular endothelial cells. Am J Physiol Cell Physiol. 2003; 285: C959–C967.
Tricarico D, Mele A, Lundquist AL, Desai RR, George AL Jr, Conte Camerino D. Hybrid assemblies of ATP-sensitive K+ channels determine their muscle-type-dependent biophysical and pharmacological properties. Proc Natl Acad Sci U S A. 2006; 103: 1118–1123.
Aguilar-Bryan L, Clement JP 4th, Gonzalez G, Kunjilwar K, Babenko A, Bryan J. Toward understanding the assembly and structure of KATP channels. Physiol Rev. 1998; 78: 227–245.
Gribble FM, Reimann F, Ashfield R, Ashcroft FM. Nucleotide modulation of pinacidil stimulation of the cloned K(ATP) channel Kir6.2/SUR2A. Mol Pharmacol. 2000; 57: 1256–1261.
Baron A, van Bever L, Monnier D, Roatti A, Baertschi AJ. A novel K(ATP) current in cultured neonatal rat atrial appendage cardiomyocytes. Circ Res. 1999; 85: 707–715.
Poitry S, van Bever L, Coppex F, Roatti A, Baertschi AJ. Differential sensitivity of atrial and ventricular K(ATP) channels to metabolic inhibition. Cardiovasc Res. 2003; 57: 468–476.
Wang H, Gauthier BR, Hagenfeldt-Johansson KA, Iezzi M, Wollheim CB Foxa2 (HNF3beta) controls multiple genes implicated in metabolism-secretion coupling of glucose-induced insulin release. J Biol Chem. 2002; 277: 17564–17570.
Labrador V, Brun C, Konig S, Roatti A, Baertschi AJ. Peptidyl-glycine alpha-amidating monooxygenase targeting and shaping of atrial secretory vesicles: inhibition by mutated N-terminal ProANP and PBA. Circ Res. 2004; 95: e98–e109.
van Bever L, Poitry S, Faure C, Norman RI, Roatti A, Baertschi AJ. Pore loop-mutated rat KIR6.1 and KIR6.2 suppress KATP current in rat cardiomyocytes. Am J Physiol Heart Circ Physiol. 2004; 287: H850–H859.
Claycomb WC, Lanson NA Jr, Stallworth BS, Egeland DB, Delcarpio JB, Bahinski A, Izzo NJ Jr. HL-1 cells: a cardiac muscle cell line that contracts and retains phenotypic characteristics of the adult cardiomyocyte. Proc Natl Acad Sci U S A. 1998; 95: 2979–2984.
Rosenblatt-Velin N, Montessuit C, Papageorgiou I, Terrand J, Lerch R. Postinfarction heart failure in rats is associated with upregulation of GLUT-1 and downregulation of genes of fatty acid metabolism. Cardiovasc Res. 2001; 52: 407–416.
Schreiber E, Matthias P, Muller MM, Schaffner W. Rapid detection of octamer binding proteins with ‘mini-extracts’, prepared from a small number of cells. Nucleic Acids Res. 1989; 7: 6419.
Brun C, Philip-Couderc P, Raggenbass M, Roatti A, Baertschi AJ. Intracellular targeting of truncated secretory peptides in the mammalian heart and brain. FASEB J. 2006; 20: 732–734.
Hellqvist M, Mahlapuu M, Samuelsson L, Enerback S, Carlsson P. Differential activation of lung-specific genes by two forkhead proteins, FREAC-1 and FREAC-2. J Biol Chem. 1996; 271: 4482–4490.
Im SS, Kwon SK, Kang SY, Kim TH, Kim HI, Hur MW, Kim KS, Ahn YH. Regulation of GLUT4 gene expression by SREBP-1c in adipocytes. Biochem J. 2006; 399: 131–139.
Bastie CC, Nahle Z, McLoughlin T, Esser K, Zhang W, Unterman T, Abumrad NA. FoxO1 stimulates fatty acid uptake and oxidation in muscle cells through CD36-dependent and -independent mechanisms. J Biol Chem. 2005; 280: 14222–14229.
Tabibiazar R, Wagner RA, Liao A, Quertermous T. Transcriptional profiling of the heart reveals chamber-specific gene expression patterns. Circ Res. 2003; 93: 1193–1201.
Loennechen JP, Stoylen A, Beisvag V, Wisloff U, Ellingsen O. Regional expression of endothelin-1, ANP, IGF-1, and LV wall stress in the infarcted rat heart. Am J Physiol Heart Circ Physiol. 2001; 280: H2902–H2910.
Melnyk P, Zhang L, Shrier A, Nattel S. Differential distribution of Kir2.1 and Kir2.3 subunits in canine atrium and ventricle. Am J Physiol Heart Circ Physiol. 2002; 283: H1123–H1133.
Dixon JE, McKinnon D. Quantitative analysis of potassium channel mRNA expression in atrial and ventricular muscle of rats. Circ Res. 1994; 75: 252–260.
Franco D, Demolombe S, Kupershmidt S, Dumaine R, Dominguez JN, Roden D, Antzelevitch C, Escande D, Moorman AF. Divergent expression of delayed rectifier K(+) channel subunits during mouse heart development. Cardiovasc Res. 2001; 52: 65–75.
Kaab S, Barth AS, Margerie D, Dugas M, Gebauer M, Zwermann L, Merk S, Pfeufer A, Steinmeyer K, Bleich M, Kreuzer E, Steinbeck G, Nabauer M. Global gene expression in human myocardium-oligonucleotide microarray analysis of regional diversity and transcriptional regulation in heart failure. J Mol Med. 2004; 82: 308–316.
Matsuo M, Tanabe K, Kioka N, Amachi T, Ueda K. Different binding properties and affinities for ATP and ADP among sulfonylurea receptor subtypes, SUR1, SUR2A, and SUR2B. J Biol Chem. 2000; 275: 28757–28763.
Yue TL, Bao W, Gu JL, Cui J, Tao L, Ma XL, Ohlstein EH, Jucker BM. Rosiglitazone treatment in Zucker diabetic fatty rats is associated with ameliorated cardiac insulin resistance and protection from ischemia/reperfusion-induced myocardial injury. Diabetes. 2005; 54: 554–562.
Chen Y, Park S, Li Y, Missov E, Hou M, Han X, Hall JL, Miller LW, Bache RJ. Alterations of gene expression in failing myocardium following left ventricular assist device support. Physiol Genomics. 2003; 14: 251–260.
Hannenhalli S, Putt ME, Gilmore JM, Wang J, Parmacek MS, Epstein JA, Morrisey EE, Margulies KB, Cappola TP. Transcriptional genomics associates FOX transcription factors with human heart failure. Circulation. 2006; 114: 1269–1276.
Skurk C, Izumiya Y, Maatz H, Razeghi P, Shiojima I, Sandri M, Sato K, Zeng L, Schiekofer S, Pimentel D, Lecker S, Taegtmeyer H, Goldberg AL, Walsh K. The FOXO3a transcription factor regulates cardiac myocyte size downstream of AKT signaling. J Biol Chem. 2005; 280: 20814–20823.
Jorgensen SB, Wojtaszewski JF, Viollet B, Andreelli F, Birk JB, Hellsten Y, Schjerling P, Vaulont S, Neufer PD, Richter EA, Pilegaard H. Effects of alpha-AMPK knockout on exercise-induced gene activation in mouse skeletal muscle. FASEB J. 2005; 19: 1146–1148.
Giatromanolaki A, Koukourakis MI, Sivridis E, Gatter KC, Harris AL, Banham AH. Loss of expression and nuclear/cytoplasmic localization of the FOXP1 forkhead transcription factor are common events in early endometrial cancer: relationship with estrogen receptors and HIF-1alpha expression. Mod Pathol. 2006; 19: 9–16.
Shen C, Nettleton D, Jiang M, Kim SK, Powell-Coffman JA. Roles of the HIF-1 hypoxia-inducible factor during hypoxia response in Caenorhabditis elegans. J Biol Chem. 2005; 280: 20580–20588.
Tang TT, Lasky LA. The forkhead transcription factor FOXO4 induces the down-regulation of hypoxia-inducible factor 1 alpha by a von Hippel-Lindau protein-independent mechanism. J Biol Chem. 2003; 278: 30125–30135.
Frescas D, Valenti L, Accili D. Nuclear trapping of the forkhead transcription factor FoxO1 via Sirt-dependent deacetylation promotes expression of glucogenetic genes. J Biol Chem. 2005; 280: 20589–20595.
Biggs WH 3rd, Meisenhelder J, Hunter T, Cavenee WK, Arden KC. Protein kinase B/Akt-mediated phosphorylation promotes nuclear exclusion of the winged helix transcription factor FKHR1. Proc Natl Acad Sci U S A. 1999; 96: 7421–7426.
Stanley WC, Recchia FA, Lopaschuk GD. Myocardial substrate metabolism in the normal and failing heart. Physiol Rev. 2005; 85: 1093–1129.