Hypoxia-inducible factor 1 is required for remote ischemic preconditioning of the heart

Significance

Remote ischemic preconditioning (IPC) is a phenomenon in which brief cycles of limb ischemia and reperfusion, which can be induced in mice by clamping and unclamping the femoral artery and in human subjects by inflating and deflating a blood pressure cuff, result in protection of the heart against a prolonged episode of ischemia and reperfusion. The underlying mechanism may involve signals transmitted by the nervous system and by secreted factors, including interleukin-10. In this study, we demonstrate that the transcriptional activator hypoxia-inducible factor 1 (HIF-1) is necessary and sufficient for remote IPC in mice and for induction of Il10 gene expression in cultured mouse myocytes. Strategies designed to induce HIF-1 activity may afford cardioprotection in patients undergoing cardiac bypass surgery.

Abstract

Both preclinical and clinical studies suggest that brief cycles of ischemia and reperfusion in the arm or leg may protect the heart against injury following prolonged coronary artery occlusion and reperfusion, a phenomenon known as remote ischemic preconditioning. Recent studies in mice indicate that increased plasma interleukin-10 (IL-10) levels play an important role in remote ischemic preconditioning induced by clamping the femoral artery for 5 min followed by 5 min of reperfusion for a total of three cycles. In this study, we demonstrate that remote ischemic preconditioning increases plasma IL-10 levels and decreases myocardial infarct size in wild-type mice but not in littermates that are heterozygous for a knockout allele at the locus encoding hypoxia-inducible factor (HIF) 1α. Injection of a recombinant adenovirus encoding a constitutively active form of HIF-1α into mouse hind limb muscle was sufficient to increase plasma IL-10 levels and decrease myocardial infarct size. Exposure of C2C12 mouse myocytes to cyclic hypoxia and reoxygenation rapidly increased levels of IL-10 mRNA, which was blocked by administration of the HIF-1 inhibitor acriflavine or by expression of short hairpin RNA targeting HIF-1α or HIF-1β. Chromatin immunoprecipitation assays demonstrated that binding of HIF-1 to the Il10 gene was induced when myocytes were subjected to cyclic hypoxia and reoxygenation. Taken together, these data indicate that HIF-1 activates Il10 gene transcription and is required for remote ischemic preconditioning.

Coronary heart disease (CHD) is the leading cause of mortality in the US population, accounting for one in every six deaths, at a rate of one death from CHD every minute (1). Coronary artery stenosis due to atherosclerotic plaques results in reduced perfusion and myocardial ischemia. Plaque rupture results in complete arterial occlusion and the death of cardiac cells (myocardial infarction; MI) due to oxygen deprivation (2). Rapid reperfusion by thrombolytic therapy or percutaneous coronary intervention is the most important clinical factor to limit infarct size (IS), while at the same time reperfusion contributes to tissue injury by increasing intracellular reactive oxygen species and Ca²⁺ levels (3, 4). Exposure of the heart to short (5-min) episodes of ischemia (I₅) and reperfusion (R₅) protects the heart against injury caused by a subsequent prolonged episode of ischemia and reperfusion (IR), a phenomenon known as ischemic preconditioning (IPC) (5).

Although IPC was shown to have a powerful protective effect in animal models, the obvious difficulties involved in subjecting the heart to direct IPC restrict its potential clinical applications. However, the discovery that an IPC stimulus applied to the circumflex coronary artery reduced the size of an MI arising from sustained occlusion of the left anterior descending artery (6) was subsequently extended by the demonstration that the heart could be protected by subjecting a remote organ or tissue, such as kidney, small intestine, liver, or limb, to short episodes of IR, a phenomenon known as remote IPC (7–10). Application of a tourniquet to the hind limb before prolonged cardiac IR reduced arrhythmias in rats and IS in pigs (9, 11). Most recently, a randomized clinical trial was conducted in which patients who received a remote IPC stimulus consisting of three cycles of I₅ and R₅ of the arm before coronary artery bypass graft surgery were found to have significantly reduced postoperative serum levels of cardiac troponin I, which is indicative of cardioprotection (12). Given the major impact of CHD on public health and the translational potential of these findings, it is not surprising that understanding the molecular mechanisms involved in cardioprotection was identified as a high-priority knowledge gap by a National Heart, Lung, and Blood Institute workshop (13).

Cardiac function was improved and IS was decreased in mice subjected to remote IPC induced by transient occlusion of the femoral artery for three cycles ([3 × I₅R₅]^fem) followed, 24 h later, by left coronary artery occlusion for 30 min and reperfusion for 120 min ([I₃₀R₁₂₀]^cor). Plasma interleukin-10 (IL-10) levels were increased >threefold in mice subjected to remote IPC (14). The cardioprotective effects of remote IPC were lost in IL-10 knockout mice and in wild-type (WT) mice treated with a monoclonal antibody against IL-10 receptor 1 (IL-10R1 MAb) that blocks IL-10 binding. IL-10 treatment of IL-10 knockout mice reduced IS by >50%, which was similar to the effect of remote IPC in WT mice (14). Perfusion of Langendorff isolated heart preparations with IL-10 immediately before [I₃₀R₁₂₀]^cor reduced left ventricle (LV) end diastolic pressure and IS and increased levels of phosphorylated AKT by >fivefold. These cardioprotective effects were lost when IL-10 was perfused with IL-10R1 MAb (14). Taken together, these results indicate that IL-10 is a critical mediator of cardioprotection induced by remote IPC in mice.

When tissue perfusion is impaired, reduced O₂ availability leads to increased activity of hypoxia-inducible factor 1 (HIF-1), which coordinates adaptive responses through the transcriptional activation of hundreds of target genes (15). HIF-1 is a heterodimer composed of HIF-1α and HIF-1β subunits (16). Hif1a^−/− mouse embryos, which are homozygous for a knockout allele at the locus encoding HIF-1α, die with cardiac malformations and vascular regression (17). Hif1a^+/− (HET) mice, which are heterozygous for the knockout allele, develop normally and are indistinguishable from WT littermates under normoxic conditions, but have impaired responses to hypoxia and ischemia, including a complete loss of cardioprotection following direct IPC (18), which also occurred after intracardiac delivery of small interfering RNA against HIF-1α in WT mice (19). Hif1a^+/− mice also manifest impaired recovery of limb perfusion following femoral artery ligation (20). In the present study, we investigated whether HIF-1 is required for remote IPC.

Results

Loss of Cardioprotection Induced by Remote IPC in Hif1a^+/− Mice.

HET mice and WT littermates were subjected to three 5-min cycles of femoral artery occlusion and reperfusion ([3 × I₅R₅]^fem) as a remote IPC stimulus or were subjected to a control sham surgery (exposure of femoral artery only); 24 h later, the mice were subjected to left coronary artery occlusion and reperfusion (Fig. 1A) followed by Evans blue dye staining of the area at risk (AAR) and triphenyltetrazolium chloride staining to distinguish infarcted (unstained/white) from viable (stained/red) cardiac tissue (Fig. 1B). In WT mice (n = 6–8), remote IPC significantly reduced IS, when analyzed as percentage of total LV area (LVA, 14 ± 2% after remote IPC compared with 35 ± 2% after sham surgery; mean ± SEM; P < 0.01) or as percentage of AAR (30 ± 3% compared with 66 ± 3%; P < 0.01) (Fig. 1C). In contrast, in Hif1a^+/− mice (n = 8–10), IS was not significantly different in littermates subjected to remote IPC vs. sham surgery, when IS was analyzed as percentage of total LVA (32 ± 2% vs. 26 ± 2%; P > 0.05) or as percentage of AAR (63 ± 2% vs. 53 ± 2%; P > 0.05) (Fig. 1C). Thus, as previously described for cardioprotection mediated by direct IPC (18), partial deficiency of HIF-1α is associated with a complete loss of cardioprotection mediated by remote IPC.

Fig. 1.

Analysis of remote IPC (RIPC) of the heart in WT and Hif1a^+/− mice. (A) Schematic of the experimental protocol is shown. WT and Hif1a^+/− (HET) littermate mice were subjected to control sham surgery (CON) or RIPC, which consisted of three 5-min cycles of femoral artery occlusion and reperfusion, followed 24 h later by left coronary artery ligation for 30 min followed by 120 min of reperfusion (I₃₀R₁₂₀). (B) Analysis of MI. The hearts were stained with Evans blue dye, to identify AAR, and triphenyltetrazolium chloride, to determine IS, and then cut into five sections. (C) IS was calculated as a percentage of total LVA or AAR. The AAR as a percentage of LVA is also shown. Data are shown as mean ± SEM (n = 6–10 in each group); **P < 0.01 vs. WT/CON; ^##P < 0.01 vs. WT/RIPC.

Loss of AKT Phosphorylation in Hearts of Hif1a^+/− Mice Subjected to Remote IPC.

Cardioprotection induced by remote IPC is mediated in part by activated AKT and inhibitors of phosphatidylinositol 3-kinase (PI3K), which activates AKT, block cardioprotection induced by remote IPC (14). AKT phosphorylation at serine-473 was significantly increased in the hearts of mice that were subjected to remote IPC compared with control mice that were subjected to sham surgery (Fig. 2A). In contrast, the remote IPC stimulus had no effect on AKT phosphorylation in the hearts of HET mice, which is consistent with the loss of cardioprotection described above (Fig. 1).

Fig. 2.

RIPC induces cardiac AKT phosphorylation and increases plasma levels of IL-10. Mice (n = 4–5 in each group) were subjected to sham surgery (CON) or three 5-min cycles of femoral artery occlusion and reperfusion (RIPC). After 24 h, heart tissue and blood plasma were collected. (A) Heart lysates were prepared and analyzed for levels of phosphorylated (P-AKT) and total (AKT) AKT protein by immunoblot assays (Lower), and the P-AKT:AKT ratio was determined for each sample (mean ± SEM; Upper). *P < 0.05 vs. WT/CON; ^#P < 0.05 vs. WT/RIPC. (B) ELISA for IL-10 was performed on plasma samples (mean ± SEM). **P < 0.01 vs. WT/CON; ^##P < 0.01 vs. WT/RIPC.

Loss of IL-10 Expression Induced by Remote IPC in Hif1a^+/− Mice.

We hypothesized that the loss of cardioprotection in HET mice was due to failure to increase circulating IL-10 levels in response to the remote IPC stimulus. Analysis of serum IL-10 levels by ELISA 24 h after sham surgery or [3 × I₅R₅]^fem revealed that in WT mice (n = 4–5), circulating IL-10 levels increased from 32 ± 8 pg/mL in controls to 119 ± 23 in mice subjected to [3 × I₅R₅]^fem (Fig. 2B). In HET mice (n = 4–5), baseline levels of IL-10 (29 ± 9 pg/mL) were similar to WT mice, but did not increase in response to the remote IPC stimulus (27 ± 3 pg/mL). These data indicate that HIF-1 is required for the increased circulating IL-10 levels induced by the remote IPC stimulus.

Activation of HIF-1 in Limb Is Sufficient to Increase Serum IL-10 Levels and Mediate Cardioprotection.

AdCA5 is a recombinant adenovirus that encodes a constitutively active form of HIF-1α that activates transcription of HIF-1 target genes even under nonhypoxic conditions (21, 22). Injection of AdCA5 into the hind limb muscle of nonischemic mice is sufficient to mobilize bone marrow angiogenic cells (20). WT mice (n = 3–4 each) received an intramuscular injection of AdCA5 or AdLacZ, which is a control adenovirus encoding Escherichia coli β-galactosidase, and serum IL-10 levels were analyzed 24 h later. The mean IL-10 levels in AdCA5-treated mice were increased to 145 ± 21 pg/mL compared with 52 ± 10 pg/mL in AdLacZ-treated mice (Fig. 3A). To determine whether AdCA5 can mediate cardioprotection similar to that induced by remote IPC, WT mice received an intramuscular injection of AdCA5 or AdLacZ into the hind limb and 24 h later were subjected to cardiac IR. IS, when measured as a percentage of the total LVA or AAR, was significantly reduced in AdCA5-treated mice (Fig. 3B). Taken together, the data presented in Figs. 1–3 indicate that HIF-1α expression in the hind limb is necessary and sufficient to increase plasma IL-10 levels and to induce remote IPC of the heart.

Fig. 3.

Effect of HIF-1α gene therapy. WT mice (n = 3 in each group) received an intramuscular injection of a replication-defective recombinant adenovirus encoding E. coli β-galactosidase (AdLacZ) or a constitutively active form of HIF-1α (AdCA5). (A) Twenty-four hours later, blood plasma was collected and ELISA for IL-10 was performed (mean ± SEM). *P < 0.05 vs. AdLacZ. (B) Twenty-four hours later, the mice were subjected to coronary artery occlusion and reperfusion and myocardial IS was determined as a percentage of the total LVA or AAR (mean ± SEM). *P < 0.05; **P < 0.01 vs. AdLacZ.

IL-10 mRNA Expression Is Induced in Myocytes Subjected to Hypoxia−Reoxygenation.

To study the molecular mechanisms involved in the induction of IL-10 expression by remote IPC, we used a cell model in which differentiated C2C12 mouse myocytes were exposed to control tissue culture conditions (CON) of 20% O₂ (95% air and 5% CO₂) or were subjected to cyclic hypoxia−reoxygenation (CHR), which consisted of hypoxia (1% O₂ for 5 min) followed by reoxygenation (20% O₂ for 5 min), repeated for a total of six cycles. CHR increased HIF-1α protein levels in differentiated C2C12 cells (Fig. 4A). Reverse transcription and quantitative real-time PCR analysis revealed that expression of IL-10 mRNA was induced fourfold by CHR, whereas Rpl13a mRNA levels were unchanged (Fig. 4B). Treatment of cells with acriflavine, which inhibits HIF-1−dependent transcription by blocking the dimerization of HIF-1α and HIF-1β (23), abolished CHR-induced IL-10 mRNA expression in C2C12 cells (Fig. 4C).

Fig. 4.

HIF-1 is required for induction of IL-10 mRNA expression in mouse myocytes subjected to CHR. (A) Differentiated C2C12 cells either were subjected to six cycles that each consisted of 5 min at 1% O₂ followed by 5 min at 20% O₂ (CHR) or were exposed to 20% O₂ continuously for 1 h (CON). Whole cell lysates were prepared, and immunoblot assays of HIF-1α and actin were performed. Representative blots from three independent experiments are shown. (B) C2C12 myocytes were exposed to CHR or CON. Reverse transcription−quantitative real-time PCR (RT-qPCR) analyses of IL-10 and Rpl13a mRNA were performed (mean ± SEM, n = 3–4). **P < 0.01 vs. CON. (C) Differentiated C2C12 cells were pretreated with the HIF-1 inhibitor acriflavine or vehicle (0.1% DMSO) for 30 min and exposed to CHR or 20% O₂. RT-qPCR analyses of IL-10 mRNA were performed (mean ± SEM, n = 3). ***P < 0.001 vs. CON/DMSO; ^#P < 0.05 vs. CHR/DMSO. (D and E) C2C12 myocytes were stably transduced with lentivirus expressing a control scrambled shRNA (shSC) or shRNA against HIF-1α (sh1α-2; D) or HIF-1β (sh1β-2; E). C2C12-shSC and C2C12-sh1α-2 (D) or C2C12-sh1β-2 (E) cells were exposed to CHR or CON. Immunoblot assays of HIF-1α or HIF-1β and actin were performed. (F and G) Differentiated C2C12-shSC and C2C12-sh1α-2 (F) or C2C12-sh1β-2 cells (G) were exposed to CHR or CON. RT-qPCR analyses of IL-10, Rpl13a, HIF-1α, or HIF-1β mRNAs were performed (mean ± SEM, n = 4). ***P < 0.001 vs. shSC/CON; ^###P < 0.001 vs. shSC/CHR.

To validate the pharmacological studies, we transduced C2C12 myocytes with lentivirus encoding one of two different short hairpin RNAs (shRNAs) targeting HIF-1α (sh1α-1 or sh1α-2) or HIF-1β (sh1β-1 or sh1β-2), or a scrambled control shRNA (shSC). Expression of sh1α-1 or sh1α-2 reduced HIF-1α mRNA in C2C12 myocytes compared with cells expressing shSC (Fig. 4F and Fig. S1C). HIF-1α protein levels were also decreased in C2C12 knockdown cells (Fig. 4D and Fig. S1A). Similarly, expression of sh1β-1 or sh1β-2 decreased HIF-1β mRNA and protein expression in C2C12 myocytes exposed to CHR or CON (Fig. 4 E and G and Fig. S1 B and D). IL-10 mRNA levels were significantly reduced in differentiated C2C12-sh1α or C2C12-sh1β cells exposed to CHR (Fig. 4 F and G and Fig. S1 C and D). In contrast, Rpl13a mRNA expression was not affected by knockdown of HIF-1α or HIF-1β in C2C12 myocytes (Fig. 4 F and G and Fig. S1 C and D).

HIF-1 Binds Directly to the Il10 Gene in Mouse Myocytes Subjected to CHR.

HIF-1 binds to the DNA sequence 5′-RCGTG-3′ (R = A or G) located near HIF-1−regulated genes (24). Inspection of the nucleotide sequence of the mouse Il10 gene revealed candidate HIF-1 binding sites in the 5′-flanking sequence (5′-FS) and 3′-FS, at nucleotides −332 and +10,861, respectively, relative to the transcription start site (Fig. 5A). To determine whether HIF-1 binds to either of these sites, differentiated C2C12 mouse myocytes were subjected to CON or CHR conditions and analyzed by chromatin immunoprecipitation (ChIP) assays. Quantitative real-time PCR revealed that the nucleotide sequence encompassing the putative HIF-1 binding site in the 3′-FS was selectively enriched after chromatin from CHR myocytes was immunoprecipitated with specific antibodies against HIF-1α or HIF-1β, but not with nonspecific IgG (Fig. 5 D and E). In contrast, no enrichment of the candidate site in the 5′-FS was detected by ChIP (Fig. 5 B and C). It should be noted that only the functional HIF-1 binding site in the 3′-FS conforms to an extended consensus sequence, 5′-RCGTGN_0–8CAC-3′ (N = any nucleotide), which has been identified in multiple target genes (25), including EPO, ALDOA, ANGPTL4, BNIP3, COX4I2, CP, CXCR3, ILK, MXI1, PDGFB, PGF, and PKM2 (Table S1), whereas the nonfunctional 5′-FS candidate site contains nine nucleotides between the 5′-GCGTG-3′ and 5′-CAC-3′ motifs (Fig. 5A). Taken together, the data in Figs. 4 and 5 indicate that HIF-1 is required for CHR-induced Il10 gene expression in mouse myocytes and that HIF-1 binding to the 3′-FS of the Il10 gene is induced when mouse myocytes are subjected to CHR.

Fig. 5.

HIF-1 binds directly to the Il10 gene in mouse myocytes subjected to CHR. (A) Nucleotide sequence and coordinates of two candidate HIF-1 binding sites located in the 5′-FS and 3′-FS of the mouse Il10 gene. The transcription initiation site (bent arrow) is designated +1. Exons (E) and introns are not drawn to scale. (B–E) Chromatin immunoprecipitation assays. Differentiated C2C12 cells were exposed to CHR or 20% O₂ (CON). Chromatin was immunoprecipitated with total IgG, anti-HIF-1α antibody (B and D), or anti-HIF-1β antibody (C and E) and analyzed by RT-qPCR (mean ± SEM, n = 3–7). *P < 0.05; **P < 0.01 vs. CON.

Discussion

In this study, we demonstrate that HIF-1 plays an essential role in cardioprotection that is induced by remote IPC. We show by loss-of-function and gain-of-function studies that HIF-1 is necessary and sufficient to induce remote IPC in vivo and that HIF-1 is required for activation of Il10 gene expression in response to CHR in vitro. Although ischemia is a complex state that is characterized by decreased delivery of O₂ and energy substrates and increased accumulation of toxic metabolites, which are all relieved upon reperfusion, hypoxia and reoxygenation are the critical stimuli that lead to increased Il10 expression, because exposure of cultured myocytes to CHR is sufficient to induce Il10 gene transcription in a HIF-1−dependent manner. Forced expression of HIF-1α in the limb was sufficient to induce increased plasma IL-10 levels and cardioprotection in the absence of any ischemia−reperfusion stimulus. Cardiac protection induced by remote IPC was associated with increased AKT phosphorylation in the heart. Neuroprotection mediated by IL-10 also involves AKT phosphorylation, and treatment with the PI3K inhibitor wortmannin blocked AKT phosphorylation and the neuroprotective effects of IL-10 (26). Direct IPC of the heart also requires PI3K → AKT signaling (27).

IL-10 may not be the only HIF-regulated factor induced by remote IPC that mediates cardioprotection. Remote IPC was recently reported to promote cardioprotection by increasing plasma levels of CXCL12/stromal-derived factor 1 (28), which is also encoded by a HIF target gene (29). We previously showed that exposure of WT mice to cycles of ambient hypoxia and reoxygenation induced cardioprotection that was associated with induction of erythropoietin expression in the kidney, representing another form of remote preconditioning, which was lost in HET mice (30). Finally, cardioprotection induced by direct IPC of the heart, which is associated with induction of genes encoding enzymes required for adenosine production, is also lost in HET mice (18, 31). Thus, HIF-1 activates multiple protective pathways in response to ischemia. Direct IPC of the heart was shown to be dependent upon functional interaction of HIF-1α with the circadian rhythm protein PER2 (32), and further studies are required to determine whether PER2 is required for HIF-1−dependent induction of IL-10 in response to [3 × I₅R₅]^fem.

Patients undergoing elective percutaneous coronary intervention or coronary artery bypass graft surgery are subjected to IR that places them at risk for myocardial injury. When these procedures are performed on an elective basis, it may be possible to use some form of remote IPC by induction of HIF-1 in the limb before cardiac surgery, either by transient adenoviral HIF-1α gene therapy, as demonstrated in our study, or by pharmacological inhibition of the prolyl hydroxylases that promote HIF-1α degradation (19). This approach may provide a more robust induction of IL-10 and possibly other protective secreted factors than can be achieved by blood pressure insufflation, which has given mixed results in prospective randomized placebo-controlled clinical trials (12, 33).

Materials and Methods

Mice.

All animal experiments were approved by the Johns Hopkins University Animal Care and Use Committee and used 3-mo-old WT and HET male littermate mice. The mice were anesthetized by intraperitoneal (IP) injection of pentobarbital (70 mg/kg). The left femoral artery was exposed at the inguinal ligament and separated from the femoral vein and nerve. A microvascular clip was used to occlude the artery for 5 min followed by release of the clip and reperfusion for 5 min, repeated for a total of three cycles. Sham surgery was identical except that the artery was not occluded. Twenty-four hours later, the mice were anesthetized with pentobarbital (70 mg/kg IP), intubated, ventilated, and maintained at 37 °C using a heating plate and rectal thermometer. A left thoracotomy was performed and a 7-0 suture was placed around the left coronary artery ∼2 mm below the left auricle. A loop was made in the suture and tightened over a piece of polyethylene tubing, resulting in arterial occlusion, as evidenced by myocardial blanching and electrocardiographic abnormalities (ST segment elevation and QRS complex widening). After 30 min, the loop was loosened to recover perfusion for 120 min. Following (I₃₀R₁₂₀)^cor, the loop around the coronary artery was retightened and 0.5% Evans blue dye was injected into the aorta. The heart was then excised, frozen, and cut transversely into five sections, each of which was weighed and stained with 1.5% triphenyltetrazolium chloride for 15 min at 37 °C. Both sides of each section were photographed and IS, AAR, and total LVA were determined by computerized planimetry [Image J; National Institutes of Health (NIH)]. IS was calculated as a percentage relative to AAR (IS/AAR × 100%) or LVA (IS/LVA × 100%).

Cardiac Phospho-AKT and Plasma IL-10 Assays.

Mice were subjected to remote IPC or sham surgery as described above. Twenty-four hours later, the mice were anesthetized and blood was drawn by cardiac puncture into a heparinized syringe for plasma isolation and aliquots were subjected to IL-10 ELISA using a commercial kit (BioLegend). The heart was excised, whole tissue lysates were prepared, and aliquots were subjected to immunoblot assay using antibodies specific for phospho-AKT (Ser-473) and total AKT (Cell Signaling Technology).

Adenoviral Injection.

WT mice received an intramuscular injection of 2 × 10⁸ plaque-forming units of AdLacZ or AdCA5, divided among three sites in the adductor muscle and one site in the gastrocnemius muscle of the left hind limb as previously described (20). Twenty-four hours after adenoviral injection, the mice were subjected to coronary artery occlusion and reperfusion as described above.

Cell Culture and CHR.

C2C12 cells were cultured in DMEM with 10% heat-inactivated FBS at 37 °C in a 5% CO₂/95% air incubator. Myocyte differentiation was initiated by switching the cells to DMEM with 2% heat-inactivated horse serum for 7 d. For hypoxic exposure, the C2C12 differentiation media was replaced with DMEM/2% horse serum that was preequilibrated with 1% O₂ and placed in a modular incubator chamber (Billups-Rothenberg), which was flushed with a gas mixture containing 1% O₂, 5% CO₂, and 94% N₂. The CHR protocol consisted of six cycles of 1% O₂ for 5 min, followed by 20% O₂ for 5 min at 37 °C. Cells were pretreated with HIF-1 inhibitor acriflavine (5 μM) or vehicle alone (0.1% DMSO) for 30 min before CHR.

Plasmid Constructs and Virus Production.

The shRNA oligonucleotides targeting scrambled control, or mouse HIF-1α or HIF-1β (Table S2) were annealed and ligated into AgeI/EcoRI-linearized pLKO lentiviral vector as previously described (34). Plasmid constructs were confirmed by nucleotide sequencing. Lentivirus encoding shSC, sh1α, or sh1β was generated by transfection of HEK293T cells with transducing and packaging vectors pMD2.G and psPAX2 (Addgene). After 48 h, virus particles in the medium were harvested, filtered, and transduced into C2C12 cells.

Immunoblot Assays.

C2C12 cells were lysed in radioimmunoprecipitation assay buffer. Aliquots of lysate were mixed with Laemmli buffer, boiled for 5 min, subjected to 7% SDS/PAGE, and transferred to a nitrocellulose membrane. The membrane was blocked and incubated overnight with anti-HIF-1α (Santa Cruz), anti-HIF-1β or anti-Actin (Novus Biologicals) primary antibody at 4 °C, followed by goat anti-rabbit or goat anti-mouse IgG conjugated to horseradish peroxidase for 1 h at room temperature. After washing, the immune complexes were detected by chemiluminescence using ECL Prime (GE Healthcare).

RNA Analysis.

Total RNA was isolated using TRIzol (Invitrogen) and reverse transcription−quantitative PCR assays were performed as previously described (35) using primers with nucleotide sequences that are listed in Table S2. Levels of IL-10, HIF-1α, HIF-1β, or Rpl13a mRNA were normalized to the levels of 18S rRNA in the same sample.

ChIP Assays.

C2C12 cells were exposed to CHR or 20% O₂, cross-linked with 1% formaldehyde for 20 min at 37 °C, and quenched in 0.125 M glycine. DNA was purified after IP of sonicated cell lysates, or was purified directly from input lysates, and quantified by SYBR Green Real-time PCR (Bio-Rad). The following antibodies were used: HIF-1α (Santa Cruz) and HIF-1β and control IgG (Novus Biologicals). The primers used for quantitative real-time PCR are shown in Table S2. Fold enrichment was calculated based on the cycle threshold (Ct) as 2^−Δ(ΔCt), where ΔCt = Ct_IP − Ct_Input and Δ(ΔCt) = ΔCt_antibody − ΔCt_IgG.

Statistical Analysis.

Data are expressed as mean ± SEM. Differences between experimental groups were analyzed by Student's t test or two-way ANOVA with Tukey’s post hoc test. P < 0.05 was considered significant.