Hypoxia-inducible factor (HIF) plays an important role in regulating gene expression in response to ischemia. Although activation of HIF-1 in muscle tissue was found during ischemia in vivo, the meaning and mechanisms in isolated cells are still incompletely understood. We studied activation of HIF-1 in skeletal muscle cells cultured in either their undifferentiated myoblast state or differentiated into myotubes. HIF-1 was activated in myoblasts and myotubes by hypoxia and simulated ischemia. Induction of adrenomedullin mRNA and, to a lesser extent, VEGF mRNA correlated well with the induction of HIF-1α protein in both cell types. Enzymes of glycolysis-like lactate dehydrogenase and pyruvate kinase showed upregulation of their mRNA only under hypoxic conditions but not during simulated ischemia. Phosphofructokinase mRNA showed no significant upregulation at all. Although HIF-1 was activated in myotubes during simulated ischemia, myotubes died preceded by a loss of ATP. Myoblasts survived simulated ischemia with no decrease in ATP or ATP turnover. Furthermore, pharmacological inhibition of HIF-1 hydroxylases by dimethyloxalylglycine (DMOG) increased HIF-1α accumulation and significantly upregulated the expression of adrenomedullin, VEGF, lactate dehydrogenase, and pyruvate kinase in myoblasts and myotubes. However, DMOG provided no protection from cell death. Our data indicate that HIF-1, although activated in myotubes during simulated ischemia, cannot protect against the loss of ATP and cell viability. In contrast, myoblasts survive ischemia and thus may play an important role during regeneration and HIF-1-induced revascularization.
- vascular endothelial growth factor
- lactate dehydrogenase
acute ischemia of muscle tissue can result from arterial embolism, acute atherosclerotic thrombosis, and prolonged arterial clamping during surgery but also may be due to traumatic injury and vessel destruction. During ischemia, with the decrease of oxygen (O2) and substrate concentrations, the synthesis of ATP is impaired. Although ATP utilization in skeletal muscle may be adapted to reduced perfusion (16), a mismatch of ATP need to ATP production will decrease the tissue ATP content and impair the activity of enzymes requiring ATP, such as ion transporters. A disturbed ion homeostasis will cause cell edema and unspecific activation of proteases (8). Although skeletal muscle tissue is relatively resistant to ischemia, viability decreases after 4 h of complete interruption of blood supply (8, 14). Long-lasting ischemia results in cell death and inflammation, later followed by repair processes including formation of new capillaries and new muscle fibers (32).
When hypoxia persists for more than a few minutes, cells respond by changes in gene expression. Hypoxia-inducible factor-1 (HIF-1) acts as a global coordinator of this cellular response to regain O2 homeostasis (35). HIF-1 is a heterodimer of an α- and β-subunit that are both members of the family of basic-helix-loop-helix (bHLH)/PER, ARNT, SIM (PAS) transcription factors (10). Abundance and activity of HIF-1α and the homologous HIF-2α are instantaneously increased upon hypoxia (18), whereas steady-state levels of the HIF-1β protein, also known as aryl hydrocarbon nuclear translocator (ARNT), are not affected by changes in oxygen tension (10). The oxygen sensors that control the abundance of HIF-α proteins in the cells are a family of novel prolyl hydroxylases named PHD1, PHD2, and PHD3 (10). In the presence of oxygen, these PHDs catalyze the Fe(II)-dependent hydroxylation of specific prolyl residues within the oxygen-dependent degradation (ODD) domains of HIF-α subunits (10). Once hydroxylated, the von Hippel-Lindau tumor suppressor protein, which is the recognition component of an E3 ubiquitin ligase complex, binds to the HIF-α subunits and thereby links prolyl hydroxylation to ubiquitination and proteasomal degradation (10). In response to oxygen deprivation, the α-subunits become stable and both subunits heterodimerize in the nucleus. There, dimer specifically binds to target gene motifs called hypoxia response elements (HREs) to induce hypoxia-inducible gene expression (5). HIF-1 target genes can be grouped by function. The first group comprises genes to increase blood oxygen capacity such as erythropoietin and transferrin. A second group contains genes for proangiogenic proteins such as vascular endothelial growth factor (VEGF) and adrenomedullin, and a third group has genes associated with glycolysis and glucose uptake. In addition, a variety of different genes encoding for growth factors were lately described as HIF-1 targets (5).
HIF-1α expression was found to be prominent in mouse skeletal muscle in vivo even under normal physiological conditions (39). Overexpression of stabilized HIF-1α leads to an increase in vessel formation without an increase in vascular leakiness, as shown for overexpression of VEGF alone (30, 42). In addition, from in vivo studies, an increase in HIF-1α and VEGF mRNA in human and mouse muscle tissue after ischemia was reported (13, 20, 31, 41). Interestingly, in one study, increased HIF-1α and VEGF in ischemic muscle tissue was exclusively localized to endothelial cells (15). Although HIF-1 activation is believed to increase glycolytic activity to improve ATP supply during hypoxia (34), so far HIF-1 activation in muscle tissue during ischemia has only been associated with adaptive vascularization.
To further investigate the question whether HIF-1 is able to protect muscle cells during ischemia by upregulation of glycolytic enzymes, we used the C2C12 mouse myoblast cell line as a culture model. C2C12 myoblasts can be differentiated in vitro to myotubes with properties of mature skeletal muscle [e.g., contractibility, creatine kinase, and myosin heavy chain expression (22, 24)]. With this model, the response of myotubes toward hypoxia and simulated ischemia was studied and compared with the responses of their precursor myoblasts. mRNA expression of the HIF-1 target genes that immediately adapt cells to hypoxic stress, namely, phosphofructokinase (PFK), pyruvate kinase (PK), and lactate dehydrogenase (LDH), was determined as well as expression of adrenomedullin (ADM) and VEGF because of their angioproliferative capacity.
MATERIAL AND METHODS
Dulbecco's modified Eagle's medium (DMEM), sodium pyruvate, penicillin, streptomycin, and Oligofectamine were purchased from Invitrogen (Karlsruhe, Germany), and fetal calf serum was obtained from PAA Laboratories (Cölbe, Germany). Heat-inactivated horse serum, insulin-transferrin-sodium-selenite medium supplement, 1-(4,5-dimethylthiazol-2-yl)-3,5-diphenylformazan (MTT), Triton X-100, NaCl, MgCl2, EDTA, HEPES, PMSF, DTT, Na3VO4, glycerol, horseradish peroxidase-conjugated goat anti-rabbit IgG, perchloric acid, P1,P5-di-(adenosine-5′)pentaphosphate, creatine kinase, 3,5-dichloro-2-hydroxybenzenesulfonic acid, MOPS, lactate oxidase, and firefly lantern extract were obtained from Sigma-Aldrich (Taufkirchen, Germany). The ECL Advanced Western blotting system was obtained from Amersham Biosciences (Freiburg, Germany), the commercial protein assay reagents and acrylamide were obtained from Bio-Rad Laboratories (Munich, Germany), and SDS was obtained from Roth (Karlsruhe, Germany). Dimethyloxalylglycine (DMOG) was purchased from Alexis Deutschland (Grünberg, Germany), and the Protean nitrocellulose membranes were obtained from Schleicher & Schuell (Dassel, Germany). The qPCR Mastermix for SYBR green was obtained from Eurogentec (Seraing, Belgium). The anti-HIF-1α antibody was obtained from Novus Biologicals (Acris, Hiddenhausen, Germany). The oligo(dT), Tfx-20 reagent, and Moloney murine leukemia virus (MMLV) reverse transcriptase were purchased from Promega (Mannheim, Germany), and the firefly luciferase assay kit was obtained from Biotium (Biotrend, Cologne, Germany). Glycine, 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one, and potassium carbonate were obtained from Merck (Darmstadt, Germany), and ATP, ADP, AMP, and horseradish peroxidase were obtained from Roche Molecular Biochemicals (Mannheim, Germany). HIF-1α small interference (si)RNA (siGENOME SMARTpool M-040638-00) and nontargeting siRNA (siCONTROL nontargeting siRNA no. 2 D-001210-02) were purchased from Dharmacon (Thermo Fisher Scientific, Lafayette, CO).
Cell culture and incubation procedure.
The mouse myoblast cell line C2C12 (ATCC LGC Promochem, Wesel, Germany) was grown in DMEM supplemented with 20% fetal calf serum, 0.11 mg/ml sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified atmosphere of 21% O2-71% N2-8% CO2 (by volume). Myoblasts were cultured to ∼80% of cell density before experimental usage. Myotubes were differentiated from confluent myoblast cultures by replacement of fetal calf serum with 2% heat-inactivated horse serum for 6 days. Hypoxic incubation was performed in a hypoxic workstation with 1% O2-94% N2-5% CO2 (Invivo2 400, Ruskin Technology, Leeds U.K.). Control and hypoxic conditions were established using glucose-rich DMEM (25 mM glucose) and low-glucose and ischemic conditions were established using glucose-free medium with serum (0.1–0.6 mM glucose) at 21 or 1% O2, respectively. DMOG (1 mM), a known inhibitor of the prolyl hydroxylase enzyme family (6), was added directly at the beginning of the experiment.
Western blot analysis.
Nuclear protein extracts were prepared from 100-mm dishes of cells that were ∼95% confluent. Cells were lysed in 100 μl of modified extraction buffer previously described by Ameln et al. [20 mM HEPES, 0.2 mM EDTA, 1.5 mM MgCl2, 100 mM NaCl, 2 mM DTT, 1 mM Na3VO4, and 0.4 mM PMSF (2)], incubated for 20 min on ice, and centrifuged (8,000 rpm at 4°C for 8 min). The supernatant was used as the cytosolic fraction, and the pellet was resuspended and lysed in a second buffer (20 mM HEPES, 0.2 mM EDTA, 1.5 mM MgCl2, 450 mM NaCl, 2 mM DTT, 1 mM Na3VO4, 0.4 mM PMSF, and 20% glycerol) using a magnetic stirrer for 30 min on ice. Extracts were centrifuged (13,200 rpm at 4°C for 10 min), and the supernatant was used as the nuclear fraction. The protein concentration was determined with a commercial protein assay. Per lane, 30 μg from the nuclear extract were loaded onto a 7.5% SDS-polyacrylamide gel and, after electrophoresis, blotted onto nitrocellulose membranes. Rabbit polyclonal anti-HIF-1α (1:750 dilution) served as the primary antibody, and horseradish peroxidase-conjugated goat anti-rabbit IgG (1:1,000,000 dilution) was used as the secondary antibody. The ECL Advanced Western blotting system was used for detection. The unspecific band above the HIF-1α-specific band was used as loading control.
Reporter gene assay.
The luciferase reporter gene plasmid pH3SVL containing a SV40 promoter-luciferase unit and six HIF-1 binding sites from the transferrin enhancer was a kind gift of R. Wenger (University of Zurich, Switzerland). Myoblasts were chemically transfected with Tfx-20 transfection reagent (4.5 μl/μg plasmid DNA) for 1 h in serum-free DMEM. After overnight incubation, cells were incubated in normal cell culture medium under normoxia (21% O2; control), hypoxia (1% O2), or in the presence of 1 mM DMOG. Cells were lysed by adding 100 μl of 1× reporter lysis buffer. Luciferase activity was measured with a commercial luciferase assay kit using a Luminometer TD 20/20 (Turner Designs, Sunnyvale, CA). Luciferase activity was expressed in arbitrary units and normalized to total cellular protein.
Reverse transcription and quantitative real-time PCR.
Total RNA was extracted using the phenol-chloroform method from six-well dishes, and 1 μg of total RNA was reverse transcribed into cDNA with oligo(dT) and MMLV reverse transcriptase as described previously (9). Subsequently, cDNAs for ADM, VEGF, LDH, PK, and PFK were quantified by real-time PCR using the qPCR Mastermix for SYBR green I and the GeneAmp5700 sequence detection system (PE Biosystems, Foster City, CA). The PCR reactions were set up in a final volume of 25 μl with 0.5 μl of cDNA in 1× reaction buffer with SYBR green I, 10 pmol of forward primer, and 10 pmol of reverse primer. The primers used were as follows: HIF-1α, gaa atg gcc cag tga gaa aa (forward) and ctt cca cgt tgc tga ctt ga (reverse); HIF-2α, cct gca gcc tca gtg tat ca (forward) and gtg tgg ctt gaa cag gga tt (reverse); ADM, cgc agt tcc gaa aga agt gg (forward) and cca gtt gtg ttc tgc tcg tcc (reverse); VEGF, cca agg cca gca aaa tcc ctg tgg gcc (forward) and ccg cct cgg ctt gtc aca (reverse); PFK, gcg atc tcc agg tga atg tt (forward) and cac gtt ctt cct gct gtc aa (revers); PK, cga tct gtg gag atg ctg aa (forward) and aat ggg atc aga tgc aaa gc (reverse); and LDH, agg ctc ccc aga aca aga tt (forward) and tct cgc cct tga gtt tgt ct (reverse). The PCR amplification profile was as follows: 10 min at 95°C, followed by 40–48 cycles with 15 s at 95°C and 1 min at 60°C. Tenfold dilutions of purified PCR products starting at 1 pg to 0.001 fg were used as standards. All samples were quantified in triplicate from RNA from three separate culture dishes.
Myoblasts were grown in six-well plates to 40% confluence. Cells were transfected with DMEM without antibiotics containing 2% serum supplement instead of 20% fetal calf serum, 1% Oligofectamine, and 0.05 μM of either HIF-1α siRNA (siGENOME SMARTpool) or nontargeting siRNA (siCONTROL nontargeting) for 24 h. Afterwards, experiments were performed exactly as described above for experiments without siRNA. LDH release assays (see Cell viability) were performed to exclude toxic effects of the transfection procedure.
Cell viability was assessed by determination of LDH release and with the MTT assay. Extracellular, i.e., released, LDH activity was determined in aliquots of myoblast and myotube medium after 4, 8, 12, and 24 h of incubation in the presence and absence of glucose at 21 or 1% O2, respectively, with or without addition of DMOG using a standard assay (4). At the end of the incubation period, cellular LDH activity was determined after lysis of the cells with Triton X-100 (1% in PBS). LDH values were corrected for changes in the volume of incubation medium resulting from repetitive sampling, and released LDH activity was given as a percentage of total LDH activity.
MTT reduction capacity was assessed after 24 h of incubation in the presence and absence of glucose at 21 or 1% O2, respectively, with or without addition of DMOG. Cells were incubated for 2 h in the presence of MTT and lysed, and reduced MTT was photometrically quantified. Cell viability was calculated as a percentage of control. C2C12 myoblasts were counted after detachment with trypsin in a Neubauer chamber.
ATP and ATP turnover.
After 6 h of incubation, myoblasts or myotubes grown on 100-mm cell culture dishes were covered with 2.5 ml of ice-cold perchloric acid (1 M). Cell remnants were scraped off the dish, and the suspension was subsequently neutralized with ice-cold potassium carbonate solution. After 15 min of incubation on ice, precipitated potassium perchlorate was removed by centrifugation.
ATP was determined from the supernatant using the luciferin/luciferase chemiluminescence assay: 500 μl of freshly prepared buffer (75 mM glycine, 15 mM MgCl2), 10 μl of firefly lantern extract (5 mg/ml), and 10 μl of the sample were mixed in a luminescence vial (Sarstedt, Nümbrecht, Germany) that was subsequently placed inside a Lumat LB 9507 luminometer (Berthold, Bad Wildbad, Germany). Chemiluminescence counts were detected over a period of 15 s. The concentration of ATP in the samples was calculated from the standard curve.
ATP turnover was determined after 5 and 6 h of incubation, respectively, by measuring ATP, phosphocreatine, and lactate concentrations from cell supernatants and cellular lysates. Samples from cell supernatants were mixed with an equal volume of perchloric acid (2 M) and subsequently processed as described above. Thereafter, cellular lysates from the respective incubations were prepared, and ATP was determined from the lysates as outlined above. To detect phosphocreatine, 40 μl of a solution containing ADP (1 mM), AMP (2 mM), P1,P5-di(adenosine-5′)pentaphosphate (20 μM), and 10 μl of creatine kinase (350 U/ml) were premixed with 10 μl of the lysate sample and subsequently measured by chemiluminescence as described above. The concentration of ATP in the samples (constituting the added concentration of ADP and phosphocreatine) was obtained from the standard curve, and the phosphocreatine content of the samples was calculated by subtracting the “pure” ATP value.
Lactate was determined from cell lysates as well as from cell supernatants. The sample (125 μl) was mixed with 875 μl of a solution containing 3,5-dichloro-2-hydroxybenzenesulfonic acid (3.8 mM), MOPS (45 mM), KCl (45 mM), 4-amino-2,3-dimethyl-1-phenyl-3-pyrazolin-5-one (3 mM), lactate oxidase (0.5 U/ml), and horseradish peroxidase (0.23 U/ml). After incubation for 15 min in the dark, absorption was measured spectrophotometrically at 546 nm against a blank. The concentration of lactate in the samples was calculated from the standard curve.
Total protein content was calculated from parallel incubations with a commercial protein assay using bovine serum albumin as a standard. Anaerobic ATP turnover was calculated according to Spriet (37) as ATP turnover = ΔPCr + 1.5Δlactate + 2ΔATP from the amounts of ATP, phosphocreatine (PCr), and (intracellular and released) lactate determined after 5 and 6 h of the respective incubations. Values are expressed as nanomoles of ATP per milligram of protein per hour.
Viability tests and ATP determination were performed in duplicate, and all experiments were repeated at least three times. RNA experiments were performed in triplicate and were performed at least twice. Data are means ± SE. Statistically significant differences were calculated after analysis of variance (ANOVA) and Bartlett's test were performed. P values are given in legends.
HIF activation by hypoxia and simulated ischemia.
We first studied the accumulation of HIF-1α in myoblasts and differentiated myotubes under hypoxic (1% O2) or simulated ischemic conditions (1% O2 in the absence of glucose). Hypoxia and simulated ischemia significantly increased HIF-1α protein levels in myoblasts after 2 h of incubation (Fig. 1A), whereas the absence of glucose alone did not affect HIF-1α levels. In myotubes, HIF-1α levels moderately increased in the absence of glucose but increased much more strongly under hypoxia. In contrast, HIF-1α levels under simulated ischemia were lower than under hypoxia alone. By direct comparison, HIF-1α accumulation under hypoxia was higher in myoblasts than in myotubes. In myoblasts and myotubes, HIF-1α protein levels under simulated ischemia peaked at 6 h, whereas the ischemic accumulation in myotubes remained lower compared with that under hypoxia alone at all time points (data not shown).
Furthermore, we compared the expression of HIF-1α and HIF-2α mRNA in myoblasts and myotubes (Fig. 1B). Levels of HIF-1α mRNA did not change upon differentiation from myoblasts to myotubes; hence, differences in protein level for HIF-1α (Fig. 1A) were not due to different mRNA expression. In contrast, HIF-2α mRNA increased almost 10-fold upon differentiation from myoblasts to myotubes. Interestingly, absolute levels of HIF-2α were about three orders of magnitude lower than those for HIF-1α. Both HIF-1α and HIF-2α mRNA expressions were not significantly altered by hypoxia or simulated ischemia (data not shown).
HIF activation by DMOG.
To determine to what extent the accumulation of HIF-1α due to hypoxia-dependent reduction of PHD activity contributed to the cellular response, we used DMOG as a specific inhibitor of PHDs. DMOG was not toxic to the cells at concentrations of 1 mM (data not shown; see also ⇓⇓⇓Fig. 5) and induced a significant increase in HIF-1α protein in myoblasts and myotubes under normoxic conditions that was comparable to the hypoxic induction (Fig. 1C). Hypoxia and DMOG significantly increased the transcriptional activity of the reporter vector containing the HRE of the transferrin promoter in myoblasts (Fig. 1D). Repeated attempts to transiently transfect myotubes with the reporter gene construct failed. DMOG-induced luciferase expression was as strong as hypoxic induction and correlated well with DMOG and hypoxia-induced HIF-1α levels (Fig. 1, C and D).
In both cell types DMOG strongly increased the mRNA expression of the HIF-1 target genes ADM and VEGF (Fig. 2, A and B). The induction of PK and LDH expression was stronger in myoblasts than in myotubes (Fig. 2, D and E), which corresponded well with the hypoxic induction of LDH and PK described below. Induction of all four target genes was in myoblasts still significant after 24 h of incubation (data not shown). PFK mRNA was upregulated by DMOG treatment only in myoblasts after 4 h of incubation (Fig. 2C). Of note, ADM and PFK expression was increased by differentiation, whereas PK and LDH expression was higher in myoblasts. VEGF mRNA expression was not significantly affected by differentiation.
HIF target gene regulation by hypoxia and simulated ischemia.
HIF-1 target gene expression was also studied in myoblasts and myotubes under hypoxia and simulated ischemia (Fig. 3). In myoblasts as well as in myotubes, both conditions strongly induced ADM and VEGF mRNA. In contrast, expression of PK and LDH mRNA was only increased under hypoxic conditions but not under simulated ischemia. Both genes were induced to a lesser extent in myotubes than in myoblasts. PFK mRNA was affected by neither hypoxia nor ischemia in myoblasts and was even downregulated during simulated ischemia in myotubes. In contrast to myoblasts, myotubes showed a slight but significant increase in ADM, VEGF, and LDH mRNA expression after incubation in the absence of glucose alone (data not shown).
HIF-1α knockdown by siRNA.
Transfection of myoblasts with HIF-1α siRNA almost completely inhibited HIF-1α protein accumulation under hypoxia and simulated ischemia compared with nontargeting control-transfected cells (Fig. 4A). Interestingly, transfection with nontargeting controls slightly increased HIF-1α protein levels under all conditions compared with nontransfected cells (data not shown). LDH viability assays revealed that the knockdown of HIF-1α did not alter survival of myoblasts under hypoxic or simulated ischemic conditions up to 24 h (data not shown).
ADM induction under hypoxia and especially under simulated ischemic conditions was strongly reduced in HIF-1α siRNA-transfected cells, indicating an important role of HIF-1 for ADM regulation in myoblasts (Fig. 4B). Of note, absolute ADM mRNA levels were much lower under serum-free conditions, which were required for efficient siRNA treatment, compared with the experiments performed in the presence of serum (Fig. 3). However, a strong response to hypoxia and simulated ischemia was retained (Fig. 4B). In contrast, the response of VEGF, LDH, and PK toward hypoxia and simulated ischemia was lost under serum-free conditions in the nontargeting-transfected cells and was not influenced by siRNA against HIF-1α (data not shown). This lack of HIF-target gene induction under serum-free conditions has also been observed in hematopoietic cells, where inducibility depends on growth factors within the serum (25). Nevertheless the induction by DMOG strongly indicated VEGF, PK, and LDH as HIF-1 target genes in mouse muscle cells.
Cell viability under hypoxia and simulated ischemia.
When myoblasts were incubated for 24 h under control, hypoxic, or simulated ischemic conditions, no cell death was observed as determined by the lack of LDH release into the culture supernatant (Fig. 5A). In contrast, cell death significantly increased in myotubes under ischemic conditions from 8 h onward, with almost 70% dead cells after 24 h (Fig. 5B). Light microscopy observations suggested a necrotic cell death, since no nuclear condensation or bleb formation, which are typical of apoptotic cell death, were observed. In addition, myotubes were detached and had lost their usually fiberlike shape. The remaining 30% viable cells represented the amount of still nondifferentiated myoblasts within each myotube culture. Hypoxia or the absence of glucose alone (data not shown) did not increase LDH release. Both myoblast and myotube cultures showed a significant decrease in metabolic activity under ischemic conditions as determined by their capacity for MTT reduction (Fig. 5C). Hypoxia or the absence of glucose alone (data not shown) did not significantly affect MTT reduction in either myoblasts or myotubes. In myotubes, the ∼70% lower MTT reduction rate correlated well with the LDH release shown in Fig. 5B.
Because the decrease in MTT reduction in myoblasts did not correspond to the lack of LDH release and since MTT reduction depends on cell density (12), cell numbers per area were determined. Experiments were started with 1.4 × 104 cells/cm2 under all conditions. Cell density increased under control and hypoxic conditions to ∼4 × 104 cells/cm2 within 24 h. In contrast, cell numbers only moderately increased under simulated ischemia to ∼2 × 104 cells/cm2. Thus the decrease in reductive capacity (MTT assay) as observed in myoblasts was caused by a significantly lower cell number due to a decreased growth rate and not by a decreased viability.
ATP levels remained unchanged in myoblasts under control, hypoxic, and simulated ischemic conditions (Fig. 6A). In contrast, ATP levels in myotubes were only kept constant under hypoxia but significantly decreased under simulated ischemia. Incubation in the presence of DMOG did not restore ATP levels in myotubes (data not shown). We also determined the ATP turnover in myotubes compared with myoblasts under hypoxic and simulated ischemic conditions to verify the decrease in ATP generation in myotubes. ATP turnover was unaffected by simulated ischemia compared with hypoxia in myoblasts, whereas myotubes showed a significant decrease under simulated ischemia (Fig. 6B). Thus myotubes were specifically sensitive to ischemic conditions and died preceded by a decrease in ATP generation. DMOG treatment to induce HIF-1 target genes and ameliorate the ATP loss in myotubes under ischemic conditions significantly increased HIF-1-dependent gene expression but increased neither ATP levels nor myotube survival.
A hallmark of muscle ischemia caused under various conditions is tissue hypoxia, which activates HIF-1 as the coordinator of a genetic response to compensate the lack of O2 (35). In this study we used the mouse skeletal muscle cell line C2C12 to explore the response of myoblasts, resembling satellite cells, and differentiated myotubes (24) to hypoxia and simulated ischemia. Both myoblasts and myotubes responded to hypoxia with accumulation of the O2-sensitive HIF-1α subunit (Fig. 1), which results from decreased activity of the cellular oxygen sensors, the PHDs, under hypoxia (10). In addition, pharmacological inhibition of the PHDs by DMOG likewise increased HIF-1α protein levels and HIF-1-dependent gene expression. HIF-2α, the HIF-1α ortholog, also should have been induced by DMOG, but because HIF-2α mRNA levels were ∼1,000 times lower than those of HIF-1α (Fig. 1B) and no HIF-2α protein was detectable with the currently available antibodies, we focused on HIF-1α to study the muscle cell response to hypoxia.
A major finding from this study is the fact that myoblasts survived ischemia up to 24 h much better than myotubes (Fig. 5). This is most likely due to the ability of myoblasts to maintain their ATP levels and ATP turnover even under simulated ischemia (Fig. 6). In contrast, myotubes showed a rapid reduction of both ATP levels and ATP turnover, a compromised ability to reduce MTT, and an increased LDH release and died without showing any morphological signs of apoptosis. Pharmacological induction of HIF-1α by DMOG could increase neither cellular ATP nor the survival of myotubes under simulated ischemia. This is in agreement with reports that myoblasts mainly depend on glycolytic energy production (27), whereas myotubes have an increased aerobic oxidation of fatty acids (33). In addition, higher LDH and PK expression in our cells, which was downregulated during differentiation into myotubes, may also reflect this phenotypic difference. In addition, we found a higher production rate of lactate under normoxic conditions in myoblasts (5.07 μM lactate·unit LDH−1·24 h−1) than in myotubes (3.34 μM lactate·unit LDH−1·24 h−1). Interestingly, this higher rate of glycolysis in myoblasts appears to be sufficient to maintain ATP levels even under conditions of simulated ischemia (very low glucose concentration of 0.1–0.6 mM introduced with fetal calf serum in the medium).
Exercise studies with mice bearing a tissue-specific HIF-1α knockout in skeletal muscle cells indicated that HIF-1 controls the expression of enzymes of the anaerobic energy production (26). In our experiments, hypoxia caused some moderate induction of PK and LDH in myoblasts and myotubes, but no induction was observed under simulated ischemia in either cell type. PFK was not induced at any condition. Although there are reports of enhanced enzyme activity after ischemia in vivo (44), no induction of enzyme mRNAs in skeletal muscle was reported so far. It has been reported for PK isoenzymes that the HIF-1 complex and the glucose-responsive transcription factors compete for binding to DNA (21). Thus binding of yet unknown factors at the HRE may impede HIF-1-dependent gene activation during simulated ischemia, i.e., in the absence of glucose and oxygen. The observed induction of the glycolytic enzymes by pharmacological inhibition of PHDs with DMOG is not in contrast with this hypothesis, since DMOG was applied in the presence of glucose. These findings indicate that HIF-1 may not be the principal regulator of glycolytic enzyme expression in muscle cells under ischemic conditions.
Induction of HIF-1α protein under simulated ischemia was reduced in myotubes but not in myoblasts compared with induction under hypoxia alone. Lowered HIF-1α levels have also been reported for some tumor cells after glucose deprivation under simulated ischemia (23, 43). Additional experiments in one of these studies revealed impaired translation of mRNAs under glucose/oxygen deprivation (23). Thus little HIF-1α accumulation may indicate an insufficient ability of myotubes to maintain protein synthesis, whereas myoblasts preserved their cellular ATP required for de novo protein HIF-1α synthesis (36).
Interestingly, hypoxia and simulated ischemia caused a significant upregulation of genes required for vascularization, namely, ADM and VEGF. Several biological functions have been linked to ADM such as vasodilation, vascularization, and antiapoptotic effects in endothelial cells (29). In cardiac muscle cells, ADM increased after hypoxia due to HIF binding and attenuated myocardial damage after ischemia-reperfusion (17). Furthermore, ADM administration enhanced the recovery of blood flow after hindlimb ischemia-reperfusion accompanied by an increased number of bone marrow-derived endothelial cells (1). Increased VEGF expression resulting from higher HIF-1 activity in vivo correlates with vascularization and may be an important component in muscle regeneration (3, 19, 28, 30, 42). In addition, VEGF was shown to enhance C2C12 myoblast migration and to reduce cell death in differentiating myotubes and ischemic skeletal muscle in vivo (11). Our siRNA experiments indicated that ADM upregulation was mainly due to HIF-1 action, although this was probably not the case for VEGF. In tumor cells, VEGF mRNA was shown to be stabilized in hypoglycemic conditions by RNA-binding proteins (38). In skeletal muscle, the RNA-binding protein Hu protein R has been identified to increase VEGF mRNA stability after exercise (40), which might play a comparable role during ischemia. Nevertheless, several in vivo studies demonstrated the upregulation of ADM, VEGF, and VEGF receptors after ischemia and reperfusion, which was followed by increased angiogenesis and myogenesis (1, 3, 11, 28, 42).
In conclusion, myoblasts cope much better with ischemia and survive. They keep their ATP levels up and thus provide a reservoir for tissue specific repair, i.e., to generate muscle cells from their satellite phenotype. In contrast, myotubes in vitro die under simulated ischemia. This is in line with studies on muscle regeneration in which necrotic destruction and lack of regeneration of the damaged muscle fibers was observed in favor of replacement with new fibers (7, 32). Both myoblasts and myotubes increase their expression of angiogenic factors to support tissue repair. Myoblast survival does not critically depend on HIF-1 activation, and pharmacological induction of HIF-1α by PHD inhibitors like DMOG does not appear to improve myotube survival but may improve regeneration due to VEGF and ADM induction.
This study was supported by a grant from the Deutsche Forschungsgemeinschaft (Fa 225/20-2) within the SPP 1151.
Present address of N. Dehne: Faculty of Medicine, Institute of Biochemistry I, Johann Wolfgang Goethe-University, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society