HIF-1 regulation of miR-29c impairs SERCA2 expression and cardiac contractility

Abstract

The principal regulator of cellular response to low oxygen is hypoxia-inducible factor (HIF)-1, which is stabilized in several forms of heart failure. Our laboratory developed a mouse strain in which a stable form of HIF-1 can be inducibly expressed in cardiomyocytes. Strikingly, these mice show a rapid decrease in cardiac contractility and a rapid loss of SERCA2 protein, which is also seen in heart failure. Interestingly, while the SERCA2 transcript decreased, it did not fully account for the observed decrease in protein. We therefore investigated whether HIF-1-regulated microRNA could impair SERCA translation. Multiple screening analyses identified the microRNA miR-29c to be substantially upregulated upon HIF-1 induction and to have complementarity to SERCA, and therefore be a potential regulator of SERCA2 expression in hypoxia. Subsequent evaluation confirmed that miR-29c reduced SERCA2 expression and Ca²⁺ reuptake. Additionally, administration of an antagonist sequence (antimir) improved cardiac contractility and SERCA2 expression in HIF transgenic mice. To extend the significance of these findings, we examined miR-29c expression in physiological hypoxia. Surprisingly, miR-29c decreased in these settings. We also treated mice with antimir before infarction to see if further suppression of miR-29c could improve cardiac function. While no improvement in contractility or SERCA2 was observed, reduction of heart size after infarction indicated that the antimir could modulate cardiac physiology. These results demonstrate that while a HIF-1-regulated microRNA, miR-29c, can reduce SERCA2 expression and contractility, additional factors in the ischemic milieu may limit these effects. Efforts to develop miRNA-based therapies will need to explore and account for these additional countervailing effects.

NEW & NOTEWORTHY Our study demonstrated hypoxia-inducible factor-1-dependent upregulation of miR-29c, which, in turn, inhibited SERCA2 expression and reduced cardiac contractility in a transgenic overexpression system. Interestingly, these results were not recapitulated in a murine myocardial infarction model. These results underscore the complexity of the pathological environment and highlight the need for therapeutic target validation in physiologically relevant models.

Listen to this article's corresponding podcast at https://ajpheart.podbean.com/e/hif1-regulates-mir-29c-and-serca2/.

INTRODUCTION

Ischemic heart disease, caused predominately by coronary atherosclerosis, is a leading cause of morbidity and mortality worldwide (11). The heart requires a large and constant supply of ATP to pump blood throughout the body. Oxidative metabolism is typically the principal source of ATP, and therefore contracting cardiomyocytes are especially sensitive to changes in oxygen availability. A key regulator of organismal response to oxygen levels is hypoxia inducible factor (HIF)-1. HIF-1 is a highly conserved transcription factor that binds to hypoxia-response elements (HREs) within the genome and regulates gene expression important for metabolism, angiogenesis, and oxygen transport (30). HIF-1 is composed of two subunits: constitutively expressed HIF-1β and oxygen-sensitive HIF-1α, the stability and activity of which is carefully controlled through O₂-dependent posttranslational modifications (18, 43).

Our laboratory has developed a tetracycline-inducible mouse model in which transgenic HIF-1α has been mutated to allow its expression under normal oxygen conditions by substitution of alanine residues at all three points of regulation by hydroxylation (Pro⁴⁰², Pro⁵⁶⁴, and Asn⁸⁰³, subsequently referred to as HIF-PPN) (16, 20). We have demonstrated that cardiomyocyte-specific overexpression of HIF-PPN leads to a series of changes in glycolysis and oxidative phosphorylation pathways that would be expected to preserve ATP production in the setting of hypoxia. Particularly striking was the rapid cardiac dilation, impairment in contractility, and correlated decrease in SERCA2 protein (4).

SERCA2 (also known as ATP2A2) is a crucial component of excitation-contraction coupling. This ATP-dependent pump removes Ca²⁺ from the cytoplasm to regenerate ion gradients between contractions (41) using substantial amounts of ATP for this function (8, 36). Only the SERCA2a isoform is expressed in cardiomyocytes, making it essential to heart function (41). In both animal models and human patients, decreased SERCA2 protein expression is commonly observed in heart failure (2, 7, 9, 12, 28). Many of our present pharmacological therapies for heart failure work in part by activating SERCA, especially those directed at adrenergic signaling (24, 48). Another exciting potential approach to improving heart function is to use gene therapy to increase myocardial SERCA2 expression (14, 19, 53).

While SERCA2 expression dramatically decreased in hearts of transgenic mice upon HIF-PPN expression, the observed reduction in transcript was not as robust (4). A number of posttranscriptional and posttranslational processes can control protein expression including microRNAs (miRNAs), which have emerged as key regulators of mRNA stability and translation (38). We thus set out to evaluate miRNA expression in the hearts of our HIF-PPN mice for potential regulators of SERCA2 translation. Several miRNAs were upregulated in HIF-PPN hearts, but miR-29c was of particular interest, as it was identified in multiple screens and had a potential binding site within the SERCA2a 3′-untranslated region (UTR). While miR-29c has previously been implicated in the regulation of cell survival and fibrosis after infarction (40, 49), its role in excitation-contraction coupling has not been explored.

Our study reveals a role for miR-29c as a regulator of SERCA2 expression that can be modulated by HIF-1 activity. In vivo data also demonstrated that administration of an antagonist sequence (antimir) to miR-29c improved cardiac contractility in hearts overexpressing HIF-PPN. Interestingly, miR-29c was downregulated in settings of physiological hypoxia. Although antimir treatment did not appear to substantially alter SERCA2 expression or cardiac contractility after myocardial infarction (MI), we did observe less increase in heart weight. Here, we establish a clear, functional relationship among HIF-1, miR-29c, and SERCA2, but our results indicate that complex interactions in the ischemic environment may diminish these effects and must be examined carefully when considering this axis for therapeutic targets.

MATERIALS AND METHODS

Mice and reagents.

All animal protocols and experiments were approved by the Institutional Animal Care and Use Committee of the University of Hawaii at Manoa. HIF-1α-PPN/tTA mice were generated as previously described (4). Briefly, transgenic mice with α-myosin heavy chain-driven expression of the tetracycline transactivator (tTA; a gift from Dr. Glen Fishman, Jackson strain 003170) were bred with a mutated HIF-1α transgenic line driven by a tetracycline response element. Mutated HIF-1α has three amino acid substitutions (Pro⁴⁰², Pro⁵⁶⁴, and Asn⁸⁰³ to Ala, denoted as HIF-1α-PPN) that allow its stable expression in a normal oxygen environment. The HIF-PPN transgene also includes a COOH-terminal hemagglutinin (HA) tag to distinguish the transgene product from the endogenous protein. Double transgenic mice (tTA/HIF-1α-PPN) were maintained on 200 μg doxycycline/ml of 2.5% sucrose-water to suppress HIF-1α-PPN expression. All animals were treated with doxycycline from conception. Doxycycline was removed from drinking water for 3–28 days to induce expression of HIF-PPN, whereas control mice continued to receive doxycycline. All experiments used 6- to 8-wk-old male mice. For MI experiments, C57BL/6 male mice between 8 and 12 wk of age (>25 g) were used. When possible, littermate controls were used.

Antibodies used in this study were purchased from the following sources: rabbit anti-mouse HIF-1α (NB100-479, Novus Biologicals), mouse anti-rabbit GAPDH (clone GAPDH-71.1, Sigma), mouse anti-SERCA2 (clone IID8, Sigma), goat anti-human SERCA2 antibody (N-19, Santa Cruz Biotechnology), mouse anti-β-actin (clone AC-74, Sigma), mouse anti-human Akt (40D4, Cell Signaling), and rabbit anti-mouse phospho-Akt (Ser⁴⁷³, Cell Signaling).

miRNA inhibitors (antimirs) were locked nucleic acid oligonucleotides purchased from Exiqon (now Qiagen). For a control, negative control A (5′-ACGTCTATACGCCCA-3′) was used. It was not possible to design a sequence to specifically block miR-29c activity alone. Therefore, the sequence used for antimir experiments (anti-29c, 5′-GATTTCAAATGGTGCT-3′) could inhibit both miR-29b and miR-29c. Antimirs were resuspended in sterile PBS for both in vitro and in vivo experiments.

Electrocardiogram analysis.

Lead I and lead II electrocardiograms were recorded continuously in sedated mice (n = 3 mice/group) using PowerLab equipment and LabChart software (AD Instruments). QT intervals were then analyzed in LabChart by measuring from the beginning of the QRS complex to the end of the T wave.

Real time-PCR miRNA array analysis.

After 3 or 7 days of HIF-PPN induction, mouse hearts were harvested and perfused with saline. RNA was isolated by flash freezing, pulverization, and TRIzol (Invitrogen). Expression profiles of the 709 best-established miRNAs were measured by real-time PCR using the Mouse miRNome Sanger miRBase microRNA Profiler Set (version 13.1, System Biosciences) and analyzed according to the manufacturer’s instructions. Samples were normalized using geometric mean of all values. Fold changes in miRNA expression were compared with control animals (HIF-PPN mice that remained on doxycycline).

Bioinformatic tools were then used to identify miRNAs that could potentially modulate SERCA2a function. Specifically, the miRANDA algorithm, TargetScan and miRTAR were applied to identify probable target miRNA sequences in the regulatory region of SERCA2 mRNA. miRNAs with mirSVR scores less than or equal to −1 were considered in greater detail. The miRNAs with predicted binding but that were not substantially dysregulated in our RT-PCR array (<2-fold dysregulation) were excluded from further analysis.

Microarray analysis.

miRNA expression was also assessed by microarray analysis. After 3 days of HIF-PPN induction, mouse hearts were harvested and the enriched miRNA fractions were isolated using a Qiagen miRNeasy mini kit according to the manufacturer’s instructions. Two candidate hearts were selected per group based on the increased heart weight-to-body weight ratio (3-day HIF-PPN vs. on doxycycline control). These RNA samples were then pooled and submitted to LC Sciences for analysis using a Mouse miRNA Array (miRMouse v.12.0). Statistical analysis performed on normalized signal values (n = 5 signal values/probe). Microarray data are available on Gene Expression Omnibus under Accession No. GSE119535.

Transcription factor-binding site analysis.

Human and mouse genomes were evaluated for conserved regions using ECR Browser (29). These sequences were then analyzed for potential HIF-1 binding sites using rVISTA (version 2.0) (23).

Cell culture.

The murine HL-1 cell line (atrial cardiomyocyte derived) was generated and provided to us by Dr. W. C. Claycomb (Louisiana State University Medical Center) and was maintained in Claycomb media as previously described (6). Human embryonic kidney (HEK)-293 cells (American Type Culture Collection) were maintained in DMEM, 5% FBS, and 100 U/ml penicillin-streptomycin. For hypoxia induction, cells were exposed to 1% O₂ for 6–24 h using an InvivO2 400 hypoxia workstation (Baker-Ruskinn). Control cells were maintained in a normal oxygen environment (21% O₂). Cells were then washed with deoxygenated PBS and harvested in RIPA buffer or RNAlater for subsequent analysis.

Real-time PCR.

To isolate total RNA, cells or frozen heart tissue were processed using a Qiagen miRNeasy mini kit according to the manufacturer’s instructions. For mRNA, a Quanta qScript cDNA synthesis kit was used for reverse transcription. Real-time PCR was performed using FastStart Universal SYBR Green Master Mix (Roche). Murine primers were either purchased from Qiagen [QuantiTect Primer Assays; carbonic anhydrase 9 (Car9), ryanodine receptor 2 (RyR2), Atp2a2 (SERCA2), and phospholamban (Pln)] or designed (A330023F24Rik, forward: 5′-CCATTACCCTGGGCTTCCTATTGA-3′ and reverse: 5′-CAAGCACAAGGTAGCGTCAACTTC-3′ and Egln1, forward: 5′-GCTGGGCAACTACAGGATAAA-3′ and reverse: 5′-ATGTCACGCATCTTCCATCTC-3′). For miRNA, reverse transcription was performed on either total RNA or enriched miRNA fractions using a miScript II RT kit (Qiagen). Real-time PCR was performed using the miScript SYBR Green PCR kit and miScript Primer assays (Qiagen). Samples were run on a 7900HT Fast Real-Time System or QuantStudio 12K Flex Real-Time PCR System (Applied Biosystems), and relative expression was calculated by ΔΔC_t (where C_t is threshold cycle) according to standard methods. For normalization, miR-26a was used for in vitro experiments and miR-103 was used for in vivo experiments. Tangerin (EHBP1L1) transcript was used to standardize mRNA samples as its abundance is similar to many of the targets and its expression has previously been shown to be unaffected by hypoxia (4).

Western blot analysis.

Protein was extracted from cells or frozen heart tissue using standard RIPA buffer plus protease inhibitors (complete mini protease inhibitor cocktail tablets, Roche). BCA assays were used to quantify protein concentrations according to the manufacturer’s protocols (Pierce). Equal amounts of lysate were run on 4–15% or 8% SDS-PAGE gels (Tris·HCl) under reducing conditions before transfer to PVDF membranes (Immobilon-FL, EMD Millipore). Membranes were then blocked with either 1% BSA or Odyssey blocking buffer (LI-COR) and probed with primary antibodies at room temperature for 1 h or 4°C overnight. Blots were then washed before incubation with appropriate secondary antibodies from Invitrogen (anti-mouse Alexa-Fluor 563) or LI-COR (IRDye 800CW donkey anti-goat IgG, IRDye 800CW donkey anti-rabbit IgG, or IRDye 680RD donkey anti-mouse IgG) for 1 h at room temperature. Blots were then washed and scanned using a Typhoon scanner (GE Lifesciences, 535-nm excitation, 560LP emission, 500PMT) or a LI-COR Odyssey IR Imaging System. Images were quantified using ImageJ (National Institutes of Health) or Image Studio densitometry analysis software (LI-COR).

HL-1 transfections and Ca²⁺ flux assay.

HL-1 cells were grown in Claycomb medium to 80% confluence. Cells were then transfected with miScript miRNA Mimic oligos (Qiagen) using Lipofectamine 2000 or Lipofectamine LTX Plus (ThermoFisher Scientific) according to the manufacturer’s instructions. In a complementary set of experiments, HL-1 cells were transfected with the HIF-PPN plasmid previously described (4) in the presence or absence of anti-29c. Cells were either harvested in RIPA buffer or used to measure Ca²⁺ flux.

For Ca²⁺ flux measurements, transfected HL-1 cells were loaded with fura-2, a Ca²⁺-sensitive dye (ThermoFisher Scientific). Ca²⁺ transients were measured using an IonOptix system that quantifies ratiometric fluorescence (340-to-380-nm ratio) of fura-2. Analysis focused on the decay of the Ca²⁺ transient signal, and a cumulative average of 400 individual Ca²⁺ transients from each of 18 different cells was calculated. The resulting function over time, F(t), was fit to an exponential decay of the following form: F(t) = F_oexp(−t/τ), where F_o is the amplitude of the Ca²⁺ peak and t is time. The fit to this equation was optimized by a least-squares algorithm, and the magnitude of the time constant (τ) was calculated for all groups. A smaller τ is indicative of faster Ca²⁺ reuptake.

Luciferase assays.

Luciferase reporter plasmids were used to assess miRNA function. These plasmids contained a Gaussia luciferase expression construct, the 3′-UTR of the target gene [murine SERCA2a or collagen type III-α₁ (COL3A1)] and simian virus 40 promoter (GeneCopoeia). To account for potential variance in transfection efficiency, the plasmids also contain a secreted alkaline phosphatase (SEAP) reporter driven by a separate promoter (cytomegalovirus). Both reporter proteins are secreted into the cell culture media, allowing fast, simple validation of miRNA-target gene interactions by measuring the ratio of luciferase to SEAP activity in the culture media. HEK-293 cells were then transfected with the reporter plasmid and mature miRNA mimics [miR-29c-3p, miR-25-3p, and cel-miR-39-3p (negative control), Exiqon] using Lipofectamine 2000. Media were changed 24 h after transfection and then collected 24 h later to measure reporter activity (48 h posttransfection). Luciferase and SEAP activity were measured and analyzed using the Ready-To-Glo secreted Luciferase reporter assay and Great EscAPe SEAP chemiluminescence kit 2.0 (Clontech) according to the manufacturer’s protocol.

To confirm specific binding of miRNAs to the SERCA2a 3′-UTR, the predicted binding site of miR-29c (and miR-25) was mutated. The seed sequence was changed from TGGTGC to GATGTG using the Stratagene QuikChange II site-directed mutagenesis kit (Agilent) according to the manufacturer’s instructions using the following primers: 5′-GCAGTAGACAGATGTGATGTGAATACAAGTATTGTGTGC-3′ and 5′-GCACACAATACTTGTATTCACATCACATCTGTCTACTGC-3′. Successful mutation was confirmed by Sanger sequencing. Additional experiments included transfection of luciferase reporter, miRNA mimic, and an antimir to miR-29c (Exiqon) to abrogate miRNA activity.

Echo analysis.

Left ventricular (LV) function was assessed as previously described (46). Briefly, transthoracic echocardiography was performed on sentient mice using a Vevo 770 or 2100 system (VisualSonics). B-mode and M-mode imaging were used to obtain LV parasternal short-axis views. M-mode images of at least three heart beats were then used to measure fractional shortening (FS; in %) using the following equation: FS = [(LVEDD − LVESD)/LVEDD] × 100, where LVEDD is LV end-diastolic internal diameter and LVESD is LV end-systolic internal diameter.

Antimir treatment in HIF-PPN mice.

Upon doxycycline removal, anti-29c (25 mg/kg) was injected intravenously every 10 days. Sterile PBS was used as a negative control. Echocardiograms were performed at the designated time points. A subset of mice was euthanized by CO₂ after 10 days of treatment for interim assessment while the remainder were treated for the entire 28-day time course. For all mice, hearts were excised and perfused with PBS before flash freezing in liquid nitrogen for RNA and protein isolation.

MI and antimir treatment.

Permanent left anterior descending artery (LAD) ligation was performed as previously described (46). Briefly, mice were intubated for anesthesia with isoflurane. After a left lateral thoracotomy between the third and fourth rib, the pericardial sac was opened and the LAD was ligated with 7.0 silk. Ischemia was confirmed by ST segment elevation on electrocardiogram (Power Laboratory, AD Instruments) and blanching of the tissue. After the incision was closed in two layers, animals were extubated and monitored until awake and mobile. Control sham surgeries were performed identically except for ligation of the artery. At the designated time points, mice were euthanized by CO₂ before hearts were removed and perfused with PBS. Tissue was then dissected into ischemic and remainder fractions and flash frozen in liquid nitrogen for RNA and protein isolation.

For antimir treatment, mice were given 25 mg/kg negative control A antimir (Exiqon) or anti-29c by intravenous or intraperitoneal injection 1 day before surgery. Seven days after surgery, echocardiograms were performed before mice were euthanized by CO₂ and heart tissue was collected as described above.

Statistical analysis.

For two-group comparisons, data were analyzed using a Student’s t-test. For multiple groups, one-way ANOVA was used followed by pairwise multiple comparisons using the Tukey test or the Fisher least-significant-difference method. Error bars indicate SD unless otherwise indicated. P < 0.05 was considered statistically significant.

RESULTS

Myocardial HIF-PPN expression alters electrophysiology.

In addition to rapid loss of cardiac contractility in HIF-PPN-expressing hearts (4), we also observed changes in electrophysiology. The QT interval, a measure of ventricular depolarization and repolarization, increased substantially in HIF-PPN mice over time (Fig. 1). The inactivation kinetics of the L-type Ca²⁺ current, a determining factor in the duration of the QT interval, are dependent both on the transmembrane electrical potential and on the intracellular Ca²⁺ concentration (37). While we did not observe transcriptional changes in genes associated with ion channels that govern membrane potential (delayed rectifier K⁺ current and Na⁺ current), the increased QT interval was not entirely surprising given the substantial loss of SERCA2 expression and reduced Ca²⁺ reuptake we have previously documented (4).

Fig. 1.

Progressive increase in QT interval with hypoxia inducible factor (HIF)-1-PPN expression. Shown are measurements of the QT interval by electrocardiogram over a 28-day time course. Mice with increased HIF-PPN expression (upon removal of doxycycline) exhibited ECG waveforms indicative of slowed inactivation kinetics in the L-type Ca²⁺ current compared with control mice (on doxycycline). n = 3 mice/group. *P < 0.05 by Student’s t-test. Exact P values for each time point from day 0 through 28 are as follows: P = 0.5294, P = 0.1125, P = 0.0836, P = 0.0325, P = 0.0437, P = 0.0134, and P = 0.0184.

Interestingly, despite the rapid disappearance of SERCA2 protein (>80% loss within 1 wk of HIF-PPN induction), the reduction in SERCA2 transcript was modest, with a 30–50% reduction in mRNA abundance after 7 days of HIF-PPN expression (4). While a number of posttranscriptional and posttranslational processes can control protein expression, miRNAs are key regulators of mRNA stability and translation. We therefore examined whether HIF-PPN also affected miRNA expression that could inhibit SERCA2 translation.

HIF-PPN causes miRNA dysregulation.

We measured miRNA expression in HIF-PPN mice induced for 3 or 7 days using a real-time PCR array and identified 11 miRNAs that were more abundant (>2-fold) after 3 days and 16 that were more abundant by 7 days (Fig. 2A). Six of these miRNAs were upregulated greater than twofold at both time points (miR-22, miR-24, miR-27, miR-29c, miR-30c, and miR-451). To corroborate these results, we also performed microarray analysis of miRNA expression in HIF-PPN hearts that had been induced for 3 days (Table 1). Strikingly, all but one of the upregulated miRNAs detected by microarray (log₂ fold change > 1) were also observed with the RT-PCR array. The only exception was miR-30e, a closely related family member to miR-30a, miR-30c, and miR-30d, which were all upregulated in the RT-PCR array.

Fig. 2.

miR-29c regulation of SERCA2 expression. A: differentially expressed miRNAs with log₂ fold changes of >2 at 3 and/or 7 days of hypoxia-inducible factor (HIF)-1-PPN induction [off doxycycline (dox)] as identified by RT-PCR array. Fold changes are shown relative to control mice (on dox) and includes n = 3 mice/group. The dashed line indicates baseline for comparison (on dox). Student’s t-test was used to compare 3 or 7 days off dox to on dox controls. Error bars indicate 95% confidence intervals. B: Western blot for SERCA2 of HL-1 cells transfected with miR-30c, miR-29c, or miR-22. GAPDH was used as a loading control and for normalization. Quantification of blots is at the bottom compared with control. C: sequences illustrating miR-25 and miR-29c interactions with SERCA2a 3′-untranslated region (UTR). Nucleotides that were mutated to disrupt miRNA binding in the luciferase reporter assay are highlighted in gray. Lines indicate nucleotide pairings. D and E: luciferase reporter assays to measure miR-29c binding to SERCA2a 3′-UTR. D: a mutated SERCA2a 3′-UTR construct (SERCA2a mut) was used to confirm specificity of interaction. Overall P values for SERCA2a, SERCA2a mut, and collagen type III-α₁ (COL3A1) plasmids were P = 0.0009, P = 0.4082, and P = 0.0089, respectively. E: antagonist sequence (antimir) to miR-29c (anti-29c) alleviated miR-29c inhibition of SERCA2a-driven luciferase activity. Overall P values for SERCA2a + miR-29c, SERCA2a + miR-25, and COL3A1 + miR-29c were P = 0.0773, P = 0.043, and P < 0.0001, respectively. For D and E, SERCA2a + miR-25 and COL3A1 + miR-29c pairings were used as positive controls. NC, negative control. For B, D, and E, one-way ANOVA with Tukey’s multiple comparisons (SERCA2a plasmids) or Student’s t-test (COL3A1 plasmid) was used. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 for between-group comparisons. Data are representative of at least 3 independent experiments.

Table 1.

Differentially expressed miRNAs from microarray analysis

Upregulated		Downregulated
miRNA	Log2 fold change	miRNA	Log2 fold change
mmu-miR-29c	3.23	mmu-miR-155	−1.01
mmu-miR-451	2.62	mmu-miR-26b	−1.04
mmu-miR-30e	2.24	mmu-let-7e	−1.06
mmu-miR-29a	1.91	mmu-miR-345–5p	−1.38
mmu-miR-328	1.61	mmu-miR-1895	−1.39
mmu-miR-99a	1.19	mmu-miR-705	−1.42
mmu-miR-22	1.18	mmu-miR-181b	−1.47
mmu-miR-342–3p	1.16	mmu-miR-188–5p	−1.47
mmu-miR-143	1.02	mmu-miR-1224	−1.64
mmu-miR-125a-5p	1.00	mmu-miR-98	−1.67
		mmu-miR-181d	−1.81
		mmu-miR-1894–3p	−1.86
		mmu-miR-1892	−2.02
		mmu-miR-124	−2.21
		mmu-miR-574–5p	−2.53
		mmu-miR-1187	−2.55
		mmu-miR-689	−3.32
		mmu-miR-712*	−3.63
		mmu-miR-1193	−4.31

Log₂ fold changes ≥ 1 for samples from 3-day hypoxia-inducible factor-1-PPN induced mice (off of doxycycline) compared with control hearts (on doxycycline). All dysregulated microRNAs were P < 0.01 by Student’s t-test.

miR-29c inhibits SERCA2 expression.

We then used bioinformatic analysis to determine which of the upregulated miRNAs could affect SERCA2 expression and found three miRNAs with potential binding sites in the 3′-UTR of SERCA2a (miR-22, miR-29c and miR-30c). To determine which of these miRNAs were functional, we transfected HL-1 cells (atrial cardiomyocyte cell line) with each of the three miRNAs and examined SERCA2 expression. Interestingly, only miR-29c significantly downregulated SERCA2 protein levels (Fig. 2B). Luciferase reporter assays also confirmed the miR-29c-SERCA2a interaction. Since miR-29c-COL3A1 and miR-25-SERCA2a interactions have been previously demonstrated (33, 42), these were used as positive controls. miR-25 and miR-29c also conveniently have overlapping binding sites on the 3′-UTR of SERCA2a, so we mutated this region to test the specificity of the interaction (Fig. 2C). As expected, as shown in Fig. 2D, there were decreases in SERCA2a- and COL3A1-driven luciferase activity in the presence of miR-25 and miR-29c, respectively. Mutation of the SERCA2a-binding site also relieved inhibition of luciferase activity by miR-25 (Fig. 2D). Importantly, the addition of miR-29c reduced SERCA2a-driven luciferase activity by 20–30% compared with the negative control, whereas mutation of the miR-29c binding site restored luciferase levels back to baseline (Fig. 2D). We confirmed these results using antimirs to block miRNA activity. As determined in the mutagenesis experiments, inhibition of SERCA2a luciferase activity by miR-29c was reversed by its antimir (Fig. 2E). miR-29c-COL3A1 and miR-25-SERCA2a interactions also showed similar results (Fig. 2E).

HIF-PPN drives miR-29c expression and regulates SERCA2 function.

Once we established that miR-29c was upregulated in HIF-PPN hearts and inhibited SERCA2 expression, we sought evidence that HIF-1 directly regulated miR-29c expression. miR-29c, along with miR-29b, is transcribed from the noncoding RNA gene A330023F24Rik, so we examined the genomic sequence and surrounding regions for potential HREs to which HIF-1 could bind. We found one potential HRE within 100 bp of the 3′-end of miR-29c in the mouse genome, but it was not evolutionarily conserved. However, two HREs were identified in the promoter of the host gene A330023F24Rik (Fig. 3A). These response elements are also conserved in humans, increasing the likelihood that these regions are functional HIF-1-binding sites.

Fig. 3.

Hypoxia-inducible factor (HIF)-1 regulation of miR-29c and SERCA2 expression in HIF-PPN hearts. A: sequence alignment and conservation analysis for promoter region of miR-29c host gene A330023F24Rik in mouse and human genomes using rVISTA. Hypoxia-response elements (HREs) were identified using HIF-1_Q3 sequence from the TRANSFAC database. HREs are highlighted in gray. B: real-time PCR analysis of miR-29c abundance with removal/addition of doxycycline (dox; HIF-PPN induction) over a 28-day time course. n = 3 mice/group. The overall P value was P < 0.0001 using one-way ANOVA with Tukey’s multiple comparisons. ****P < 0.0001 for the 7 days off dox group compared with all other groups. C: miR-29c host gene (A330023F24Rik) expression analysis by RT-PCR. For A and C, fold changes relative to control mice (on dox) are shown and include n = 3 mice/group. D: miR-29c expression in HIF-PPN-transfected HL-1 cells measured by RT-PCR. Fold changes are shown compared with control; n = 3 mice/group. P = 0.26 by a Student’s t-test. E: Western blot for SERCA2 of HL-1 cells transfected with HIF-PPN in the absence or presence of anti-29c. GAPDH was used as a loading control and for normalization. Quantification of blots is at the bottom compared with control. One-way ANOVA with Tukey’s multiple comparisons was used. ***P < 0.001 and ****P < 0.0001 for between-group comparisons. F: measurements of Ca²⁺ transients in HL-1 cells transfected with HIF-PPN in the absence (triangles, dashed line) or presence of anti-29c (squares, dotted line) compared with control cells (circles, solid line). Analysis focused on Ca²⁺ reuptake with data displayed as the 340-to-380-nm fluorescence (F340:380) ratio over time.

We then examined miR-29c expression over time with the removal and addition of doxycycline to assess if HIF-PPN could temporally regulate its abundance. We have previously shown a rapid loss of SERCA2 protein with HIF-PPN induction (4). Protein levels were then restored upon suppression of HIF-PPN expression, which correlated with restoration of contractile function (4). This pattern was also observed with miR-29c expression. Upon removal of doxycycline, miR-29c was substantially increased in HIF-PPN hearts after 7 days (7 days off doxycycline, Fig. 3B). Reintroduction of the drug to the drinking water returned miR-29c expression to baseline (on doxycycline) within 7 days, and it remained low as long as doxycycline was present (Fig. 3B). Expression of the miR-29c host gene A330023F24Rik was also upregulated in HIF-PPN hearts off doxycycline for 3 or 7 days (Fig. 3C). These data suggest HIF-1 may directly regulate miR-29c expression through binding to HREs in its host gene.

We further explored the link among HIF-1, miR-29c, and SERCA2 through in vitro experiments. Transfection of HIF-PPN into HL-1 cells led to increased miR-29c expression and significantly reduced SERCA2 protein, but the addition of an antimir to miR-29c restored SERCA expression (Fig. 3, D and E). Additional functional data measuring Ca²⁺ reuptake confirmed the roles of HIF-1 and miR-29c on SERCA2. During the relaxation phase of excitation-contraction coupling, cytoplasmic Ca²⁺ declines as SERCA2 pumps it back into the sarcoplasmic reticulum. We measured the end of Ca²⁺ transients in HL-1 cells when reuptake occurs by monitoring the ratio of Ca²⁺-dependent fluorescent dyes. Figure 3F shows Ca²⁺ uptake was impaired in HIF-PPN-transfected cells compared with control cells, but the addition of anti-29c ameliorated this effect. We then fit the 340-to-380-nm fluorescence ratio data to a single exponential decay and calculated the associated τ of Ca²⁺ reuptake, with a smaller τ being indicative of faster Ca²⁺ reuptake. Expression of HIF-PPN correlated with a larger τ (226.5 ± 4.5 ms) and lower peak fluorescence compared with control cells (218.0 ± 3.3 ms), whereas cells cotransfected with anti-29c had a smaller τ than control cells (207.7 ± 3.3 ms). Since Ca²⁺ reuptake is a process largely dominated by SERCA2 in the myocardium (7), these data suggest that the HIF-1-mediated decrease in SERCA2 protein leads to reduced Ca²⁺ reuptake and is at least partially mediated by miR-29c.

Inhibition of miR-29c improves cardiac contractility and SERCA2 expression in HIF-PPN mice.

After demonstrating that the relationship among HIF-1, miR-29c, and SERCA2 had functional significance in vitro, we examined whether these connections could be recapitulated in vivo. We evaluated whether miR-29c contributed to the cardiac phenotype observed in our HIF-PPN mice. While control-treated mice still had substantial loss of cardiac contractility over time, administration of an antimir to miR-29c ameliorated this loss (Fig. 4A). Importantly, SERCA2 expression was also higher in antimir-treated mice compared with control mice (Fig. 4B), suggesting the improvement in heart function was due to at least partial recovery of SERCA2 expression.

Fig. 4.

Administration of anti-29c improved cardiac contractility in hypoxia-inducible factor (HIF)-1-PPN mice. A: fractional shortening (FS) measurements in HIF-PPN-induced mice [off doxycycline (dox)] given anti-29c for 28 days. PBS was used for control injections; n = 13 mice/group for first 10 days and then n = 8 for the remaining time points. The overall P value for control group was P < 0.0001 using one-way ANOVA with Dunnett’s multiple comparisons. *P < 0.05 and ****P < 0.0001 for comparisons within control group to day 0. The overall P value for the anti-29c group was P < 0.0001 using one-way ANOVA with Dunnett’s multiple comparisons. ***P < 0.001 and ****P < 0.0001 for comparisons within the anti-29c group to day 0. ^aP < 0.001 for between-group comparisons by a Student’s t-test. B: Western blot for SERCA2 in HIF-PPN hearts after 10 days anti-29c treatment. GAPDH was used as a loading control and for normalization. *P < 0.05 by a Student’s t-test. Quantification of blots is shown at the bottom; n = 4 mice/group.

SERCA2 and miR-29c expression in physiological hypoxia.

Our transgenic model is a powerful tool to examine HIF-1-specific effects, but the level of HIF activity anticipated with uninhibitable HIF-PPN is probably beyond what is normally seen even in pathological ischemia. We therefore tested whether the HIF-1-miR-29c-SERCA2 axis was relevant in nontransgenic pathophysiological models. We first assessed HL-1 cells exposed to hypoxia and found that hypoxia-induced increased HIF-1 correlated with increased expression of HIF-1 target genes (Egln1 and Car9) and decreased expression of excitation-contraction genes (SERCA2, RyR2, and Pln; Fig. 5, A and B). SERCA2 protein also significantly decreased over time upon exposure to low oxygen conditions (Fig. 5C). These findings all agreed with data observed in our HIF-PPN experiments. However, surprisingly, miR-29c expression decreased with hypoxia treatment (Fig. 5D). To determine if this was a potential artifact of our in vitro system, we also measured miR-29c expression in an in vivo setting of pathological hypoxia. We used ligation of the LAD to mimic MI and evaluated cardiac tissue 3 days after surgery. Similar to the in vitro experiments, we also found miR-29c expression was decreased in the infarcted tissue compared with the same region of sham-operated hearts (Fig. 5E). miRNA expression was not affected in the regions remote from the infarct with either treatment (Fig. 5E).

Fig. 5.

SERCA2 and miR-29c expression in physiological hypoxia. A–D: analysis of HL-1 cells exposed to 1% O₂ for up to 24 h compared with normal oxygen conditions (N). A: Western blot of hypoxia-inducible factor (HIF)-1α expression. β-Actin was used as a loading control. B: real-time PCR analysis of Ca²⁺ handling and HIF-1 target genes. Fold changes are shown compared with normoxia. With the use of one-way ANOVA with Tukey’s multiple comparisons, the overall P values for Egln1, carbonic anhydrase 9 (Car9), SERCA2, ryanodine receptor 2 (RyR2), and phospholamban (Pln) were P = 0.0011, P = 0.0004, P = 0.0014, P = 0.033, and P = 0.0076, respectively. *P < 0.05, **P < 0.01, and ***P < 0.001 for between-group comparisons. C: Western blot of SERCA2 expression. β-Actin was used as a loading control and for normalization. Right: quantification of SERCA2 protein with normal oxygen (N) set as 100%. D: real-time PCR analysis of miR-29c expression. For A–D, all experiments were performed in triplicate. For B and D, fold changes are shown compared with normal oxygen conditions. E: miR-29c expression in mouse hearts 3 days after myocardial infarction (MI). Ischemic free wall (I) and remote (R) fractions of infarcted hearts were compared with similar tissue from sham-operated mice. n = 8 sham and n = 6 MI. Fold changes are shown compared with sham I values. For C–E, the overall P values using one-way ANOVA were P = 0.08, P = 0.08, and P = 0.0078. For C and E, *P < 0.05 by a Student’s t-test.

Inhibition of miR-29c after MI.

Although miR-29c was not upregulated after MI, detectable miRNA remained and therefore could have a biological effect. Multiple studies have found decreased SERCA2 function after ischemia (12, 28) and in heart failure (2, 7, 9), so we tested whether antimir treatment could improve cardiac function and SERCA2 expression after infarction. We measured miRNA abundance 7 days after infarction and found miR-29c remained lower in the infarct region of negative control-treated mice compared with sham and remote tissue (Fig. 6A). We also verified the efficacy and specificity of our antimir by demonstrating miR-29c abundance was further decreased in all antimir-treated mice (Fig. 6A). In contrast, while the antisense sequence miR-29c-5p was also downregulated in infarcted tissue, administration of the antimir had no effect on its expression (Fig. 6A).

Fig. 6.

Anti-29c treatment in myocardial infarction (MI) mice. A: analysis of microRNA expression in negative control (NC) and antagonist sequence [antimir (anti-29c)]-treated hearts by RT-PCR. Fold changes are shown compared with sham ischemic free wall (I) NC values; n = 4 NC sham, n = 5 NC MI, n = 3 anti-29c sham, and n = 6 anti-29c MI. The overall P values for miR-29c and miR-29c-5p were both P < 0.0001 using one-way ANOVA with Tukey’s multiple comparisons. For miR-29c, lines indicate between-group comparison. Bracket denotes P < 0.0001 for all antimir-treated groups compared with NC groups. ***P < 0.001 and ****P < 0.0001 for between-group comparisons. For mir-29c-5p, #P < 0.01 compared with all other groups. B and C: heart function as assessed by fractional shortening (FS; B) and heart weight-to-body weight ratios (HW/BW; C) 7 days post-MI; n = 4 NC sham, n = 5 NC MI, n = 3 anti-29c sham, and n = 9 anti-29c MI. The overall P values for B and C as assessed by one-way ANOVA with Tukey’s multiple comparisons were P < 0.0001 and P = 0.0126. For B, ***P < 0.001 for between-group comparisons. For C, **P = 0.008 was assessed by a Student’s t-test, NC MI vs. anti-29c MI. D: Western blot for SERCA2 in 7 day post-MI hearts. Blots are shown for fractions from two representative hearts per condition. GAPDH was used as a loading control and for normalization. E: Western blots for phosphorylated (p-)Akt and total Akt. Blots are shown for fractions from two representative hearts per condition. Total Akt was used as a loading control and for normalization. For D and E, quantification of blots is shown at the right. The overall P values in D and E were P = 0.0065 and P = 0.4546 using one-way ANOVA with Tukey’s multiple comparisons. For D, *P < 0.05 for between-group comparisons. For E, P = 0.058 by a t-test, NC MI vs. anti-29c MI. For D and E, n = 4 NC sham, n = 5 NC MI, n = 3 anti-29c sham, and n = 6 anti-29c MI.

We also examined hearts for any functional changes. We did not find the substantial impairment of contractility in infarcted hearts to be improved by antimir treatment (Fig. 6B). However, an increase in heart weight was observed in antimir-treated MI mice when assessed by their heart weight-to-body weight ratio (Fig. 6C). We also observed a marked decrease in SERCA2 protein after MI but did not see any significant improvement with administration of anti-29c (Fig. 6D). To confirm the antimir was still functionally active, we also assessed phosphorylation of Akt (Ser⁴⁷³), which has been shown to increase in anti-29c-treated mice subjected to myocardial ischemia-reperfusion (49). In agreement with the published findings, we also observed an increase in phospho-Akt in the infarcted region of antimir-treated mice compared with the negative control group, although this did not reach statistical significance (Fig. 6E).

DISCUSSION

Hypoxia is an important contributor to cardiac injury during MI and HIF-1 is a key regulator of the cellular response. We used our cardiomyocyte-specific oxygen-stable HIF-1 transgenic mice to define its role in the heart more specifically. Previously, we found HIF-PPN mice had severely reduced cardiac contractility with concurrent loss of SERCA2 protein. In this study, we extend these findings by describing how HIF-1 regulates miRNA expression, specifically miR-29c, to repress SERCA2 expression and cardiac contractility.

Other groups have shown miR-29c can regulate a number of pathways involved in heart disease. In MI, miR-29a, miR-29b, and miR-29c have been shown to be downregulated in the border zone of ischemic hearts, which corresponded with an increase in fibrosis through upregulated expression of collagens, fibrillins, and elastin (40). Others found miR-29c increased apoptosis after ischemia-reperfusion through suppression of myeloid cell leukemia/Bcl-2 family apoptosis regulator (Mcl1), while administration of pioglitazone, a peroxisome proliferator-activated receptor-γ agonist, limited this effect (49). miR-29 has also been implicated in pathological cardiac hypertrophy and fibrosis through modulation of Wnt signaling (32). Additionally, miR-29c has been associated with cardiac and renal dysfunction stemming from diabetes mellitus (3, 22). Interestingly, miR-29c may also have beneficial effects in some settings. Microarray analysis showed miR-29c increased in response to aerobic training and correlated with cardiac hypertrophy and enhanced ventricular function (34). In the kidney, miR-29c has also been shown to modulate renal fibrosis, and this outcome was regulated by HIF expression (10).

Our study adds to these findings by demonstrating HIF-1 regulation of miR-29c expression in the heart and identifying a previously undescribed role for miR-29c in the regulation of SERCA2 expression and cardiac contractility. One limitation of our study is the inescapable lack of specificity of our antimir. Since miR-29b and miR-29c only differ in sequence by one nucleotide, it is not possible to design a sequence to only inhibit miR-29c with current approaches. However, while both miRNAs regulate many of the same targets and we cannot exclude the possibility that miR-29b may have similar effects, miR-29b was not upregulated in our HIF-PPN transgenic hearts. It therefore seems more likely that miR-29c is specifically regulated by HIF-1, and we chose to focus our efforts on characterizing its role in cardiac contractility.

Since miR-29b and miR-29c are transcribed together from their host gene A330023F24Rik, it is intriguing that miR-29b and -29c have diverging patterns of expression in HIF-PPN mice. We identified HREs in the promoter region of A330023F24Rik and found increased expression of the host gene with HIF-PPN induction, so one might expect miR-29b expression to also be upregulated. miR-29b can also be transcribed in conjunction with miR-29a from a separate bicistronic transcript, so it is possible that additional layers of regulation exist to specifically regulate expression of each individual miRNA. This posttranscriptional regulation may then compensate for the upregulation of transcript by HIF-1 to limit the expression of miR-29b. HIF-1 itself may also direct expression of these regulatory molecules to precisely upregulate miRNAs of interest. miRNA processing is an active area of research, and defining a role for HIF-1 in this process is of great interest for future study.

While we examined miRNA expression in HIF-PPN hearts to identify potential candidates of SERCA2 expression, a number of other miRNAs we identified have been associated with cardiac dysfunction. Interestingly, several of these miRNAs (miR-22, miR-30e, and miR-451) show a similar pattern to miR-29c, with upregulation in HIF-PPN hearts and a loss of expression in pathological settings (13, 17, 35, 44, 51, 52). To determine if these discordant patterns in HIF-1-responsive genes were miRNA specific, we also compared mRNA expression from our HIF-PPN transgenic mice to infarcted heart tissue (3 days post-MI) (4, 46). As one would expect, we found the mRNA abundance of genes involved in cell proliferation and glycolytic metabolism to be increased in both conditions including marker of proliferation Ki-67 (MKI67), Bcl/lymphoma 6B (BCL6B), phosphofructokinase (PFKP), CAR9, glucose phosphate isomerase 1 (GPI1), and GAPDH. However, several transcripts also followed the discordant pattern observed with miR-29c expression (increased in HIF-PPN, decreased in MI) including endothelin 3 (EDN3), enolase 3 (ENO3), aldolase (ALDOA), and phosphoglycerate mutase 2 (PGAM2). One possible explanation is that transcription factors may differentially interact with the promoters of these genes to modulate gene expression based on the environment. For example, with the use of PScan analysis, early growth response (EGR) and myc proto-oncogene (MYC) binding sites were enriched in the promoters of genes upregulated in both settings; these factors are known to be upregulated in hypoxia and ischemic settings (1, 46, 47). In contrast, FOS/JUN and TEAD sites are found more often in the discordant gene set. Interestingly, HIF-1-binding sites were also more highly enriched in the discordant group than in the similarly upregulated gene set, although neither reached statistical significance. It seems likely that it is the integrated result of the combination of transcription factors present in a given setting that determines gene expression. Future study will be required to further define the specific conditions under which HIF-1 target genes are increased or repressed.

Another potential explanation for this discordance in gene expression is the unnatural physiology found in our HIF-PPN transgenic hearts. HIF-1 is typically only active when oxygen levels are low, but in our mice HIF-1 is active in a normal oxygen environment. While this allows us to study the specific effects of HIF-1, the prolonged expression of a nondegradable form of HIF-1 in our model may also affect other pathways that indirectly regulate miRNA and SERCA2 expression. For example, since the prolyl and asparaginyl hydroxylases that typically degrade HIF-1α cannot modify the mutated HIF-PPN protein, there is no homeostatic check on HIF-1 action. Expression of these hydroxylases is also upregulated by HIF-1 (26). Consequently, the continual presence of HIF-PPN may drive prolyl hydroxylase activity to supranormal levels and increase hydroxylation of other target proteins. Similarly, von Hippel-Lindau tumor suppressor typically ubiquitinates hydroxylated HIF-1α, but HIF-PPN may instead allow von Hippel-Lindau tumor suppressor to favor other target proteins for degradation such as signaling (EGF receptor, Sprouty2, and β₂-adrenergic receptor) and RNA polymerase components (RPB1 and RPB7) that could affect mRNA abundance and cardiac function (50).

It is also possible that transient expression of HIF-1 may lead to a different gene expression pattern compared with sustained HIF-1 activation and would thus have different effect on miR-29c. Often, ischemia is resolved in a clinical setting through endogenous thrombolysis or therapeutic maneuvers, so it would be interesting to evaluate the effect of miR-29c on SERCA2 in an ischemia-reperfusion model. While ischemia is transient in an ischemia-reperfusion setting, HIF-1 has been shown to be stabilized after reperfusion through a variety of mechanisms (15). We have also demonstrated that even in a permanent MI model, HIF-1 protein expression is downregulated by 3 days after MI, likely through upregulation of degradatory pathways (46). We would therefore expect miR-29c expression might be similar in both models.

HIF-1 is also only one of many factors that are activated during hypoxia to restore oxygen and blood flow to tissue, and several of these have been shown to modify HIF-1 activity. For example, inhibitory PAS domain protein (IPAS) is a splice variant of HIF-3α that can directly inhibit HIF-1 binding to HIF-1β and subsequent target gene expression (25). Additionally, modification of the epigenetic landscape has also been shown in several settings to affect HIF-1 transcriptional activity (45). There are many aspects of the cellular environment that are altered by reduced oxygen availability including impaired respiratory chain activity in mitochondria, generation of reactive oxygen species, and eventually activation of apoptosis and necrosis (27). Also, pathophysiological effects of ischemia are not limited to hypoxia. The acidosis and reduced removal of active metabolic and signaling molecules resulting from coronary ligation could all modulate the response to HIF, with eventual cumulative effects on miRNA abundance and activity.

A greater understanding of the mechanisms involved in hypoxic injury is crucial to the development of more effective therapies. A compelling reason to examine HIF-1 regulation of SERCA2 was to identify novel therapeutic targets for treating ischemic heart failure. SERCA2 is an especially attractive therapeutic target due to its essential role in excitation-contraction and previously established association with heart failure. While HIF-1 has been shown to directly reduce SERCA2 expression by modulating its promoter activity (31), we chose to focus on miRNA regulation of SERCA2 because of evidence that SERCA protein was reduced more than its mRNA and due to the potential use of antimirs as pharmaceutical reagents (39). While there is some concern about potential off-target effects of antimirs, as miRNAs tend to regulate many transcripts, technological advances to precisely target specific transcripts and organs continually improve the efficacy of this approach (5).

Other groups have also examined antimir therapy to improve heart failure and reported improvement in SERCA2 expression and contractility by inhibiting miR-25 or miR-328 (21, 42). While we did observe increased miR-328 expression in our HIF-PPN mice, it was not as strong as miR-29c and we did not observe any changes in miR-25 expression. Additionally, both of these other studies used a pressure overload model (transaortic constriction) to study heart failure, not MI. Regulation of SERCA2 expression may therefore be different in different types of injury, with slower, long-term development of injury in a pressure overload model potentially being more amenable to antimir therapy than hypoxic injury with transient stabilization of HIF. Our luciferase reporter assays (Fig. 2, D and E) also indicate that miR-29c is not as strong of an inhibitor of SERCA2 expression as miR-25, so anti-miR-29c therapy may not be as effective as anti-miR-25 to preserve contractile function. While we did not see increased contractility after antimir treatment, we did see some improvement in the heart weight-to-body weight ratio (Fig. 6C), suggesting the antimir may be modulating the response to ischemic injury.

In summary, we have shown overexpression of active HIF-1 in vivo promotes transcription of a number of miRNAs, including miR-29c. We then identified an important target for miR-29c, SERCA2a, and demonstrated its effect on cardiac function. Surprisingly, these results were not recapitulated in a physiological model of cardiac ischemia using permanent LAD ligation. It is therefore likely that the environment in which HIF-1 is activated plays an important role in directing its transcriptional response. Figure 7 shows these findings and our proposed mechanism of HIF-1, SERCA2, and miR-29c interaction. This study highlights the need for validation of results from artificial transgenic systems in physiologically relevant models of disease.

Fig. 7.

Proposed regulation of SERCA2 by hypoxia-inducible factor (HIF)-1 and miR-29c. Low oxygen induces rapid HIF-1 expression. HIF-1 directly inhibits SERCA2 expression, which, in turn, leads to reduced cardiac function [low ejection fraction (EF)]. HIF-1 can also further reduce SERCA2 abundance posttranscriptionally by upregulating miR-29c expression. However, during hypoxia, another, currently unknown, environmental or transcription factor appears to prevent HIF-1 upregulation of miR-29c and instead leads to a decrease in miR-29c abundance.

GRANTS

This work was supported by National Institutes of Health Grants R01-HL-080532 and P30-GM-103341 (to R. V. Shohet), T32-HL-115505 (to A. L. Williams), as well as P30-CA-071789 and 5G12-MD-007601 to the Microscopy and Imaging Cores, respectively. This work was also supported by a postdoctoral fellowship from the American Heart Association (to A. L. Williams).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.L.W., C.B.W. and R.V.S. conceived and designed research; A.L.W., C.B.W., K.A.M., and A.A. performed experiments; A.L.W., C.B.W., K.A.M., and A.A. analyzed data; A.L.W., C.B.W., K.A.M., and R.V.S. interpreted results of experiments; A.L.W. prepared figures; A.L.W. drafted manuscript; A.L.W., C.B.W., and R.V.S. edited and revised manuscript; A.L.W., C.B.W., K.A.M., A.A., and R.V.S. approved final version of manuscript.