Adaptive Myogenesis under Hypoxia

Abstract

Previous studies have indicated that myoblasts can differentiate and repair muscle injury after an ischemic insult. However, it is unclear how hypoxia or glucose deprivation in the ischemic microenvironment affects myoblast differentiation. We have found that myogenesis can adapt to hypoxic conditions. This adaptive mechanism is accompanied by initial inhibition of the myoD, E2A, and myogenin genes followed by resumption of their expression in an oxygen-dependent manner. The regulation of myoD transcription by hypoxia is correlated with transient deacetylation of histones associated with the myoD promoter. It is noteworthy that, unlike the differentiation of other cell types such as preadipocytes or chondroblasts, the effect of hypoxia on myogenesis is independent of HIF-1, a ubiquitous regulator of transcription under hypoxia. While myogenesis can also adapt to glucose deprivation, the combination of severe hypoxia and glucose deprivation found in an ischemic environment results in pronounced loss of myoblasts. Our studies indicate that the ischemic muscle can be repaired via the adaptive differentiation of myogenic precursors, which depends on the levels of oxygen and glucose in the ischemic microenvironment.

Skeletal muscle possesses the remarkable ability to withstand chronic ischemia. However, the newly developed muscle fibers tend to be smaller in diameter than those of healthy, nonischemic myofibers (46). Interestingly, chronic exposure to high altitude can also result in smaller myofiber area and lower muscle mass in mountaineers (25). Similar myofiber differences are also observed in healthy people who live at high altitude for generations (25). Taken together, these observations suggest that decreased oxygen tension or hypoxia alone may be sufficient to alter the differentiation of myoblasts.

Ischemic muscle damage is repaired by myogenic satellite cells, a small population of precursor cells located between myofibers (12, 23, 47). When stimulated, the otherwise quiescent satellite cells undergo terminal differentiation into myocytes to regenerate damaged myofibers. It has been found that cultured satellite cells form new muscle fibers when transplanted into ischemic muscle (12). The myogenic differentiation of myoblasts is highly orchestrated by a family of myogenic regulatory factors (MRFs), such as myoD, myogenin, myf5, and MRF4 in collaboration with the E2A gene products (E12 or E47) and/or the myogenic enhancing factors (6, 41, 53). Although a plethora of information on the embryonic development of muscle exists, there is still a lack of clear understanding of how postnatal myogenic satellite cells are regulated during muscle regeneration, especially by their microenvironment.

Tissue ischemia in myocardium and skeletal muscle results in hypoxia and elevates the expression of the hypoxia-inducible transcription factor HIF-1α and its downstream gene, VEGF (32, 40). Although ischemic tissue is exposed to hypoxia and decreased nutritional supply, cells will alter their metabolism, proliferation, and differentiation programs in response to hypoxia alone. Accumulating evidence indicates that oxygen levels modulate adipogenesis, chondrogenesis, epidermal development, myeloid differentiation, neural differentiation, and trophoblast differentiation (17, 19-21, 37, 45, 51, 55). Therefore, we hypothesized that hypoxia could also modulate myogenic differentiation.

In this study, we investigated the effects of hypoxia on myogenesis, using the mouse myoblast cell line C2C12, a widely used model to study postnatal skeletal muscle formation, as well as primary myoblasts (39, 50). The physiological partial O₂ pressure (pO₂) levels in skeletal muscle are around 35 to 40 mm Hg, or about 5% O₂ (4, 8). Since myogenesis was not affected significantly at 5% O₂, we investigated the effects of three different levels of hypoxia: physiological hypoxia at 2% O₂, pathological hypoxia at 0.5% O₂, and extreme pathological hypoxia at 0.01% O₂. The expression of myoD mRNA was transiently inhibited in C2C12 myoblasts at 2 and 0.5% O₂. Consequently, the induction of myogenin mRNA was significantly delayed in myogenically stimulated C2C12 myoblasts under hypoxia. Ectopic expression of these MRFs rescued myogenic differentiation of C2C12 myoblasts under hypoxia. Consistently, primary myoblasts that already express high levels of myogenin protein overcame the inhibitory effects of hypoxia. The regulation of myoD transcription by hypoxia is correlated with transient deacetylation of histones associated with the myoD promoter. Furthermore, myofibers formed under hypoxia appeared to be smaller than those that developed at 21% O₂. This observation provides a molecular basis to explain the contribution of hypoxia to the morphological differences found in regenerated myofibers under chronic ischemia, as well as skeletal muscle from highland dwellers.

MATERIALS AND METHODS

Plasmids.

The following retroviral constructs were made: myoD, E2A, and myogenin were cloned into pLZRS; for constitutively active HIF-1α mutants, the ΔODD (deletion of amino acids 401 to 603) and P402A/P564G mutants were cloned into pWZL. All constructs were verified by sequencing.

Cell culture, differentiation, and morphological analysis.

C2C12 mouse myoblasts (American Type Culture Collection, Manassas, Va.) were cultured in growth medium (GM) or Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% fetal bovine serum (FBS), penicillin, and streptomycin. Myogenic differentiation of myoblasts was induced at about 70% confluence in the differentiation medium (DM) or DMEM containing 2% horse serum (ES). DM was replaced every day until the end of experiments. Primary myoblasts were isolated from mouse limb skeletal muscle and cultured on a collagen-coated surface in DMEM-F10 (1:1) supplemented with 20% FBS and 2.5 ng of basic fibroblast growth factor (bFGF) per ml (39, 50). Primary myoblasts were induced to undergo myogenic differentiation in DMEM containing 5% ES. For experiments under the glucose-free condition, the glucose-free DMEM was supplemented with 1 mM sodium pyruvate, All culture media contained 20 mM HEPES, pH 7.4, to maintain the pH stability under both normoxia and hypoxia.

To evaluate the effects of hypoxia on myogenesis, myoblasts were placed in an adjustable hypoxia chamber with real-time pO₂ readout (Invivo 200; Biotrace International) and the media were replaced everyday inside the chamber. To visualize myofiber formation, cells were fixed in 2% formaldehyde in phosphate-buffered saline (PBS) and stained with hematoxylin. Phase-contrast microscopy was performed on an inverted microscope equipped with a film camera. Pictures were taken with Kodak T64 positive films.

For the evaluation of the role of the Notch signaling pathway in myogenesis under hypoxia, C2C12 cells were induced to differentiate for 3 days in DM at 21% or 0.5% O₂ in the presence of N-[N-(3,5-difluorophenacetyl)-l-alanyl]-(S)-phenylglycine t-butyl ester (DAPT) (43). The extent of myogenic differentiation was analyzed by immunofluorescence with anti-myosin heavy chain (MHC) antibody as described below.

Immunofluorescence.

C2C12 cells with or without differentiation were fixed in 2% formaldehyde in PBS, permeabilized in 0.1% Triton X-100, blocked in 5% ES-PBS-0.01% Triton X-100 for 1 h, incubated with monoclonal anti-MHC antibody (MF20 in hybridoma culture supernatant from the Developmental Studies Hybridoma Bank; 1:10 in 1% ES-PBS-0.01% Triton X-100) for 3 h, and then incubated with Alexa 448-conjugated anti-mouse immunoglobulin G (IgG) (1:500) for 1 h. Fluorescent cells were examined and documented with an inverted fluorescent microscope equipped with computer-interfaced charge-coupled device camera. All procedures took place at room temperature.

Retroviral infection.

Retroviruses were produced with the Phoenix cell system (3, 36). C2C12 cells at about 20% confluence were infected two times with retrovirus containing 8 μg of Polybrene per ml by centrifugation at 700 to 900 × g for 1 h, followed by overnight incubation at 32°C. The infected cells were induced to differentiate as described above.

Stable C2C12 cell lines with myoD promoter-driven luciferase reporter gene.

The following myoD promoter-driven luciferase reporter gene constructs cloned in pGL2 (26) (courtesy of J. P. Capone, MacMaster University, Ontario, Canada) were linearized by SalI digestion: full-length (4-kb distal enhancer plus 2.5-kb proximal promoter), 4-kb distal enhancer, 2.5-kb proximal promoter, and SV40-driven luciferase constructs. C2C12 cells were transfected with Fugene 6 (Promega) with each of the above linearized luciferase constructs. For selection of stable populations, we cotransfected C2C12 cells with pcDNA3.1-zeo (linearized by BglII digestion) at a of 1 (zeo):5 (luciferase) ratio. The stable clones were selected with zeocin at 200 μg/ml.

For luciferase reporter assays, the stable C2C12 cells were induced to differentiate in DM as mentioned above. The effects of hypoxia on reporter gene activity were investigated by placing the cells at 0.5% O₂ in the hypoxia chamber with real-time pO₂ readout (Invivo 400; Biotrace International), and the media were replaced every day inside the chamber. Cell lysates were prepared at the indicated time; the luciferase activity was measured with luciferase substrate reagents (Promega) as described previously (55). The luminescence units were normalized by total protein concentrations to adjust for differences in cell number.

Northern and Western blotting analysis.

Total cellular RNA was isolated with Trizol reagent (Life Technologies). The following plasmids were used for cDNA template preparations by restriction digest: pcDNA3/myoD, pcDNA3/E47 (E2A), pEMSV/myogenin, Id1 (IMAGE 4206508), Id2 (IMAGE 1547176), and Id3 (IMAGE 3513700). Probes were radiolabeled with [α-³²P]dCTP by random priming. Hybridization was carried out at 65°C for 6 to 12 h. The radioactive blot was visualized on a Storm 860 PhosphorImager (Molecular Dynamics, Sunnyvale, Calif.).

For Western blotting analysis, total cell lysates were prepared on ice with 25 mM HEPES buffer, pH 7.4, containing 1% NP-40, 150 mM NaCl, 2 mM EDTA, and a protease inhibitor cocktail (Complete; Boehringer Mannheim). Nuclear extracts were prepared on ice. Briefly, cells were scraped into cold PBS and lysed in buffer A [10 mM Tris, pH 8.0, containing 3 mM CaCl₂, 2 mM Mg(OAc)₂, 0.1 mM EDTA, 0.5% NP-40, 1 mM dithiothreitol (DTT), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 320 mM sucrose]. After centrifugation, the cytoplasmic fractions were collected and stored at −80°C until use. The isolated nuclei were washed in buffer A without NP-40, and solubilized first with low-salt buffer (20 mM HEPES, pH 7.9, containing 1.5 mM MgCl₂, 20 mM KCl, 0.2 mM EDTA, 1% NP-40, 0.5 mM DTT, 0.5 mM PMSF, protease inhibitor cocktail, and 25% glycerol) and then an equal volume of high-salt buffer (800 mM KCl, with the rest of the components the same as in low-salt buffer). Supernatants were stored at −80°C until use. Equal amounts (25 μg/lane) of lysates were subjected to Western blotting with polyclonal rabbit anti-myoD and antimyogenin antibodies at 1:1,000, followed by incubation with alkaline phosphatase-conjugated anti-rabbit IgG. Protein bands were visualized with ECF substrates (Amersham) on a Storm 860 PhosphorImager. β-Actin (Sigma) was used as a loading control for whole-cell lysates, as well as for verification of cytoplasmic lysates. TAFII-70 (BD Biosciences) was used as a control for nuclear extracts.

For Western blotting of acetylated histones H3 and H4, C2C12 cells were induced to differentiate in DM for up to 3 days at 21 or 0.5% O₂. Whole-cell lysates were prepared on ice in sodium dodecyl sulfate (SDS) sample buffer containing protease inhibitors. After electrophoresis (15% gel) and electrotransfer, the acetylated proteins were detected with rabbit anti-acetylated H3 or anti-acetylated H4 (Upstate Biotechnology), followed by chemoluminescence. As a positive control for histone acetylation, C2C12 cells were treated for 4 to 6 h in 5 mM sodium butyrate (NaBut). We probed actin (Sigma) for a loading control.

RT-PCR.

C2C12 cells were induced to differentiate for 3 days in DM at different pO₂ levels. Total cellular RNA was isolated with Trizol reagent (Life Technologies). A 25-μl reaction with 1 μg of total RNA was carried out with the SuperScript III First-Strand Synthesis system for reverse transcription-PCR (RT-PCR) (Invitrogen), according to the manufacturer's recommended protocol. Two microliters of the reverse-transcribed product was used as a template in a 50-μl PCR (35 cycles). The following primers were used for PCR: (i) notch1, forward, 5′-CCTTCCTAGGTGCTCTTGCG-3′, and reverse, 5′-TGCGGTCTGTCTGGTTGTGC-3′; (ii) notch2, forward, 5′-ACCCCTCCTGCTACCTGTCA-3′, and reverse, 5′-GATAGGGTCCCTTGGATGGC-3′; (iii) notch3, forward, 5′-CAAATGGAGGTCGGTGCACCC-3′, and reverse, 5′-TGGGCTGCAGCTGACACTCAT-3′; (iv) HERP2, forward, 5′-GCGGAGAGAATGGAAACTTG-3′, and reverse, 5′-GCTCAGATAACGGGCAACTTCG-3′; (v) DEC2, forward, 5′-CAGGACAGAAACCTCCAAATCG-3′, and reverse, 5′-CTCCAAATGCCCCAGTGTTG-3′; and (vi) β-actin, forward, 5′-CCACCAGACAGCACTGTGTT-3′, and reverse, 5′-ACCGAGCGTGGCTACAGCTT-3′.

ChIP.

C2C12 cells were induced to differentiate in DM for up to 3 days at 21 or 0.5% O₂. C2C12 cells were harvested at the indicated time for chromatin immunoprecipitation (ChIP) assays with polyclonal rabbit anti-acetylated histone H3 (Upstate Biotechnology), according to the manufacturer's recommended protocol. Briefly, cells were incubated for 10 min at 37°C in 1% formaldehyde and then scraped off in ice-cold PBS. Cell pellets were solubilized in SDS lysis buffer and sonicated four times for 15 s each at the maximal output with a Branson Sonifier 250. After centrifugation, the clarified supernatants were treated with protein A-agarose beads, salmon sperm DNA, and control rabbit IgG to reduce nonspecific binding. The precleared cell lysates were incubated overnight with 3 μg of polyclonal rabbit anti-acetylated H3. The immune complexes were recovered by adding protein A-agarose beads and eluted into 250 μl of elution buffer. After the cross-linking was reversed, nuclear proteins were removed by protease K digestion and genomic DNA was recovered by phenol-chloroform extraction. The precipitated DNA was dissolved in 20 μl of Tris-EDTA (TE) buffer. The myoD promoter gene was analyzed with nested primers in two consecutive rounds of PCR for 20 cycles each. For the first round PCR, 2 μl of DNA was used as a template with the following primers: 5′ primer, CCTTGGCTCAACTTCTCTGG; and 3′ primer, AGGAAGGAGGGCAGAGAGAC. For the second round PCR, 1 μl of first round PCR product was used as template with the following primers: 5′-primer, CCTTGGCTCAACTTCTCTGG; and 3′ primer, TCCTCCAGCCTGTACTGACC.

RESULTS

Myogenesis under hypoxia.

We used the C2C12 mouse myoblast cell line, the widely accepted myogenesis model, to investigate the mechanisms by which hypoxia affects myogenesis. Myoblasts are induced to differentiate upon medium change from GM (20% FBS) to DM (2% ES). Proliferating C2C12 myoblasts in GM express myoD in its inactive form. Upon the switch to DM, MyoD becomes activated, induces cell cycle arrest via p21 expression, and promotes terminal myogenic differentiation through the induction of myogenin (41). After 3 days in DM at 21% O₂, the fibroblast-like myoblasts differentiated into myocytes that fused into myofibers (Fig. 1A and B) that express MHC (myosin heavy chain, green fluorescence, Fig. 1B). In contrast, myogenesis was inhibited at ≤2% O₂, with the strongest inhibition found at 0.01% O₂ (Fig. 1A, B, and C). The hypoxic inhibition of myogenesis was further confirmed by Western blotting analysis of troponin T, another marker of myogenesis (unpublished observation). Interestingly, myofibers formed at 2% O₂ were less robust, with smaller cross-sections than their counterparts at 21% O₂ (Fig. 1A). Since the possibility existed that the apparent inhibition of myogenesis was due to cell death, we analyzed the viability of differentiating C2C12 cells at extreme hypoxia (0.01% O₂) by trypan blue exclusion (Fig. 1D). There was about 5% cell death after 1 day of extreme hypoxia treatment, ≤2% cell death after 2 days, and no detectable cell death on day 3, indicating that the induction of cell death cannot explain the decrease in mature myocytes. Similarly, a recent study has shown that hypoxia does not significantly alter the cell cycle profile of differentiating myoblasts (18). In conclusion, the effects of hypoxia on myogenesis, especially at high oxygen tensions, are independent of perturbation in cell cycle or apoptosis.

FIG. 1.

Effects of hypoxia on myogenesis. (A, B, and C) C2C12 cells were differentiated at different pO₂ levels for 3 days. Cells were fixed and stained either with hematoxylin (A) or for MHC (green) by immunofluorescence with propidium iodide (red) as a nuclear counterstain (B). Panel C represents the average number of MHC-positive nuclei from four random fields. (D) C2C12 cells were differentiated for up to 3 days at 21 or 0.01% O₂. Cellular viability was analyzed by trypan blue staining. Bar, 50 μm.

We directly investigated the ability of myoblast differentiation to adapt to hypoxia from two perspectives: (i) the reversibility of hypoxic repression of myogenesis and (ii) the effects of prolonged hypoxia on myogenesis. To address the reversibility issue, we induced myogenic differentiation at 2, 0.5, or 0.01% O₂ for 1, 2, or 3 days before transferring the cells to 21% O₂ for another 3 days of differentiation. C2C12 cells continued to differentiate even after 3 days of exposure to 2 or 0.5% O₂ (Fig. 2A). Indicative of adaptive expression, cells pretreated for 3 days at 2 or 0.5% O₂ differentiated better than those pretreated for 1 day only. However, the reversibility of myogenic differentiation was compromised at extreme hypoxia at 0.01% O₂ (Fig. 2A). Nevertheless, partial differentiation still occurred after 1 to 3 days of exposure to 0.01% O₂, although the myofibers were much smaller than those recovering from 2 or 0.5% O₂. These results demonstrate that myogenesis can recover from physiological (2% O₂) and pathological (0.5% O₂) hypoxia, but only to some extent from extreme (0.01% O₂) hypoxia. The degree of differentiation after hypoxia preincubation (Fig. 2A) reflects the levels of MRF expression at the time of reoxygenation. The different levels of MRF expression result from adaptive transcription during the hypoxia treatment (see Fig. 4A).

FIG. 2.

Adaptation of myogenesis to hypoxia. (A) C2C12 cells were differentiated at different pO₂ levels for 1 to 3 days, followed by 3 days of differentiation at 21% O₂. (B) C2C12 cells were differentiated at 0.5% O₂ for 3, 6, or 12 days continuously, with a 3-day differentiation at 21% O₂ for comparison. Myofiber formation was analyzed after hematoxylin staining (A and B). Results are representative of at least three independent experiments. Bars, 50 μm (A) and 100 μM (B).

FIG. 4.

Effects of hypoxia on myogenic gene expression and mRNA stability. (A) C2C12 cells were differentiated at different pO₂ levels for 1 to 3 days. Temporal expression of myogenic genes was analyzed by Northern blotting of total RNA. Results are representative of at least three independent experiments. (B) C2C12 cells and primary myoblasts were treated for the indicated period of time (in hours) in DM containing 5 μg of actinomycin D (ActD) per ml at 21% and 0.5% O₂, respectively. Total RNA was prepared, and Northern blotting was done as described for panel A. Relative band intensity was analyzed with NIH Image 1.63.

The adaptive myogenic differentiation of C2C12 myoblasts was also investigated by continuous exposure of differentiating myoblasts to 0.5% O₂ for up to 12 days. On day 3, myogenesis was significantly repressed at 0.5% O₂, compared to that at 21% O₂ (Fig. 2B). However, extensive myogenesis occurred by day 6 and progressed further by day 12 at 0.5% O₂ (Fig. 2B). These findings indicated that myogenesis was able to adapt to chronic hypoxia (0.5% O₂).

Myogenesis can adapt to deprivation of glucose or inhibition of mitochondrial activity.

From the perspective of energy metabolism, hypoxia represses mitochondrial respiration while promoting glycolysis (10). Glucose utilization, in turn, synergizes with hypoxia to regulate gene functions, such as c-Jun/AP-1 (30). We investigated how myogenesis was affected by the inhibition of mitochondrial activity or by glucose deprivation. NaN₃ can irreversibly inhibit cytochrome c oxidase (50% inhibitory concentration, <10 μM) in C2C12 cells (31), thus blocking the oxidative phosphorylation in mitochondria. However, myogenic differentiation of C2C12 myoblasts did not change significantly when chronically treated with 100 μM NaN₃ in the presence of 4.5 g of glucose per liter (Fig. 3A), supporting the concept that myoblasts can differentiate by anaerobic metabolism only. In contrast, myogenic differentiation was significantly delayed when myoblasts were treated with 2-deoxyglucose in the presence of 4.5 g of glucose per liter and 10 mM sodium pyruvate or deprived of glucose (Fig. 3B). Not surprisingly, more significant inhibition was found when C2C12 differentiation was carried out in glucose-free medium containing 100 μM NaN₃ (Fig. 3B). These cells were still viable but failed to differentiate. However, when exposed to hypoxia in glucose-free media, a condition mimicking severe ischemia, myoblasts rapidly lost viability and were unable to adapt (Fig. 3C). Therefore, transient repression and eventual adaptation of myogenesis are highly dependent on the severity of hypoxia and glucose deprivation. Interestingly, myoblasts retain the unique ability to adapt to either stress alone but cannot recover when exposed to severe hypoxia and glucose deprivation simultaneously, indicating that energy metabolism or glucose-dependent signaling is essential for myogenic differentiation.

FIG. 3.

Role of cellular energy metabolism in myogenesis. (A) C2C12 cells were differentiated for 3 or 6 days at 21% O₂ in the presence of 4.5 g of glucose per liter and NaN₃ (0 or 100 μM). (B) C2C12 cells were differentiated for 3 or 6 days at 21% O₂ in the presence of 4.5 g of glucose per liter and 2-deoxyglucose (0 or 100 mM) or in the presence of NaN₃ (0 or 100 μM) but without glucose. (C) C2C12 cells were differentiated for 3 or 6 days at 0.5% O₂ in glucose-free medium containing NaN₃ (0 or 100 μM). Cells were fixed and stained for MHC (green) by immunofluorescence with propidium iodide (red) as a nuclear counterstain. Bars, 100 μm (A, B) and 50 μm (C).

Hypoxia decreases myogenesis by repression of the myoD, E2A, and myogenin genes.

To understand the molecular mechanisms of hypoxic repression of myogenesis, we investigated the expression of the key MRFs during myogenic differentiation of myoblasts under hypoxia (Fig. 4A). myoD and E2A were constitutively expressed in C2C12 myoblasts, and their levels of expression did not change significantly during myogenic differentiation at 21% O₂. In contrast, the expression of the myogenin gene, a downstream target of the myoD/E2A heterodimer, increased rapidly as C2C12 myoblasts underwent differentiation at 21% O₂. Consistently, the Id genes (Id1, Id2, and Id3), inhibitors of myoD/E2A heterodimer formation in proliferating myoblasts, were rapidly down-regulated after the initiation of myogenesis. Consistent with the repression of myofiber formation by hypoxia, myoD, E2A, and myogenin gene expression decreased under hypoxia (Fig. 4A). The degree of inhibition was inversely related to pO₂ levels. Extreme hypoxia completely inhibited the expression of both the myoD and myogenin genes and strongly repressed E2A expression. It is noteworthy that myoD expression was able to recover within 48 h of exposure to 2 or 0.5% O₂. Such adaptive expression of myoD was echoed by the expression of the myogenin gene (Fig. 4A) and myofiber formation (Fig. 2B), indicating that the newly synthesized myoD was functional under hypoxia. In contrast, the myogenesis-dependent expression of Id genes was not affected significantly at either 2 or 0.5% O₂ (Fig. 4A).

To understand the nature of repressed myoD expression under hypoxia, we investigated the stability of myoD mRNA at 0.5% O₂, using actinomycin D. Our data revealed that myoD stability in C2C12 cells did not change significantly under hypoxia (Fig. 4B). Consistently, hypoxia did not significantly alter myoD stability in primary myoblasts (Fig. 4B). In contrast to C2C12 myoblasts (Fig. 4A), undifferentiated primary myoblasts also express high levels of myogenin mRNA (Fig. 4B). Nevertheless, hypoxia only slightly decreased myogenin gene stability in primary myoblasts (Fig. 4B). These findings indicate that the transient repression of myoD expression in C2C12 myoblasts results from active inhibition of myoD transcription by hypoxia.

The expression of myoD and myogenin proteins correlated with their respective mRNA expression. During differentiation at 21% O₂, the myoD levels in C2C12 cells did not change significantly, whereas the myogenin levels increased (Fig. 5A) with the accumulation of myogenin mRNA (Fig. 4A). Hypoxia caused a substantial decrease in myoD protein in C2C12 cells that was proportional to the severity of hypoxia (Fig. 5A). There was a concomitant reduction of myogenin under hypoxia. As expected, myoD was predominantly nuclear (Fig. 5B), indicating that hypoxia did not affect the subcellular localization of myoD protein. Consistent with the recovery of mRNA expression, levels of both myoD and myogenin proteins increased by day 3 at either 2 or 0.5% O₂ (Fig. 5A).

FIG. 5.

Effects of hypoxia on stability and expression of myoD protein. (A, B) C2C12 cells were differentiated at different pO₂ levels for 1 to 3 days. Temporal changes in myoD and myogenin proteins were analyzed by Western blotting of whole-cell lysates with β-actin as a control (A). (B) Cytoplasmic and nuclear lysates were prepared from differentiating C2C12 cells on days 2 and 4. Cellular distribution of myoD was analyzed by Western blotting with β-actin as a cytoplasmic control and TAFII-70 as a nuclear control. Results are representative of at least three independent experiments. (C) C2C12 cells and primary myoblasts were treated for the indicated period of time in DM containing 25 μg of cycloheximide (CHX) per ml at 21 and 0.5% O₂, respectively. Levels of myoD protein were analyzed by Western blotting. Protein loading was verified by Ponceau S staining of nitrocellulose blots, as well as by Western blotting for β-actin. Band intensity was analyzed with NIH Image 1.63. Results are representative of at least three independent experiments. (D) C2C12 cells were cultured in DM for the indicated period of time (days) at 0.5% O₂. Before lysate preparation, C2C12 cells were treated with 10 μM MG132 for 6 h. Levels of myoD protein were analyzed by Western blotting.

myoD protein is highly labile (38), with a half-life of about 40 min in C2C12 myoblasts and about 1 h in primary myoblasts (Fig. 5C). Our data indicated that pathological hypoxia (0.5% O₂) did not significantly alter the stability of myoD protein (Fig. 5C), suggesting the hypoxia-induced decrease of myoD protein is a direct consequence of the repressed myoD mRNA levels under hypoxia. This conclusion was further supported by the experiment in which we treated C2C12 myoblasts with a proteasome inhibitor, MG132, after the cells were incubated at 0.5% O₂ for 1, 2, or 3 days. There was significant decrease of myoD protein on day 1 under hypoxia. Treatment with MG132 resulted in only slight accumulation of myoD protein on day 1, suggesting repressed protein synthesis of myoD under hypoxia. Increasing amounts of myoD protein (Fig. 5A and D) occurred concomitantly with the adaptive expression of myoD mRNA (Fig. 4A) from day 1 to day 3 under hypoxia. This observation supports the concept that the transient decrease in myoD protein results from the transient repression of myoD mRNA under hypoxia.

Together, our findings suggest that hypoxia transiently inhibits myogenesis by repressing MRF expression both at the mRNA and protein levels, without significant effect on the stability of either mRNA or protein. The decrease of myoD protein directly results from the hypoxia-induced repression of myoD mRNA. Adaptive myogenesis occurs via the recovery of MRF gene transcription from repression by hypoxia.

Hypoxic inhibition of myogenesis can be reversed by ectopic expression of myoD or the myogenin gene together with E2A.

Hypoxic inhibition of myogenesis correlates with the rapid repression of the myoD, E2A, and myogenin genes within the first day. If myoD and E2A were the critical targets of hypoxia, ectopic expression of the myoD, E2A, and myogenin genes should restore myogenesis under hypoxia (Fig. 6). Ectopic expression of myoD, E2A, or the myogenin gene alone in C2C12 cells robustly enhanced myogenesis at 21% O₂, compared to that in the controls. Although myogenesis was inhibited strongly at 0.5% O₂, myoD, E2A, or the myogenin gene alone was sufficient to restore myogenesis to a level similar to that at 21% O₂. However, infection with a retrovirus containing a single-gene construct of the same genes was not able to restore myogenesis at 0.01% O₂. Double infection with retrovirus containing E2A and myoD or the myogenin gene to ensure the formation of the myogenic heterodimers myoD/E2A and myogenin/E2A further promoted myogenesis at 2 and 0.5% O₂. Even at 0.01% O₂, a substantial amount of fiber formation and MHC expression was found in cells expressing either myoD/E2A or myogenin gene/E2A (Fig. 6). These observations were consistent with the decreased transcription of MRF genes under hypoxia (Fig. 4A). These results indicate that hypoxic inhibition of myogenesis was mediated primarily by the repression of MRF expression.

FIG. 6.

Rescue of hypoxic inhibition of myogenesis by ectopic MRF expression. C2C12 cells were infected with retrovirus containing the indicated genes and differentiated at different pO₂ levels for 2 days. Cells were fixed and stained for MHC (green) with propidium iodide (red) as a nuclear counterstain. Results are representative of two independent experiments. Bar, 50 μm.

Physiological support for this conclusion came from our experiment using primary myoblasts. We found that the primary myoblasts were resistant to hypoxic repression and still capable of myofiber formation at the pathological level of hypoxia, or 0.5% O₂ (unpublished observation). In contrast to C2C12 myoblasts, the undifferentiated or proliferating primary myoblasts already express high levels of myogenin mRNA and myogenin protein (Fig. 7A). Upon myogenic stimulation under hypoxia, myoD expression decreased both at mRNA and protein levels, whereas the expression of both myogenin mRNA and its protein managed to remain at very high levels under hypoxia (Fig. 7A). This finding provides physiological support for the notion that MRF expression can overcome the inhibitory effects of hypoxia on myogenesis. Consistent with C2C12 myoblasts (Fig. 3), the primary myoblasts can also differentiate under glucose-free but normoxic conditions while maintaining high levels of myoD and myogenin proteins (Fig. 7B). When induced to differentiate under the ischemic condition (i.e., the glucose-limited and hypoxic condition), the primary myoblasts failed to form myofibers and experienced dramatic loss of both myoD and myogenin protein (Fig. 7B). This observation further supports the notion that severe ischemia is deleterious for myoblasts to regenerate and repair damaged muscle fibers.

FIG. 7.

Effects of hypoxia on the expression of the myoD and myogenin genes in primary myoblasts. (A) Primary myoblasts were cultured in differentiation medium for the indicated period of time (days) at 21 and 0.5% O₂, respectively. Changes in the expression of myoD and myogenin mRNA were investigated by Northern blotting. Levels of myoD and myogenin proteins were analyzed by Western blotting. (B) Primary myoblasts were cultured in DM without glucose for the indicated period of time (days) at 21 and 0.5% O₂, respectively. Levels of myoD and myogenin proteins were analyzed by Western blotting.

Hypoxic inhibition of myogenesis is independent of HIF-1.

The hallmark cellular response to hypoxia is the stabilization and activation of HIF-1 (9, 42). The HIF-1 pathway plays an important role in the regulation of many cellular processes, such as cell survival and apoptosis, cell cycle progression, differentiation, and tumor progression (22, 48, 54, 55). A recent study has shown that genetic deletion of HIF-1α in skeletal muscle exerts significant effects on the energy metabolism of skeletal muscle (34). To assess the role of HIF-1 in the myogenic differentiation of C2C12 myoblasts, we infected proliferating C2C12 cells with retrovirus containing an O₂-insensitive and constitutively active form of HIF-1α: ΔODD (deletion of the oxygen-dependent degradation domain) and the P402A/P564G mutant (11, 42) (Fig. 8A). These two constitutively active HIF-1α mutants exhibited robust transcription activity at 21% O₂ as evaluated by the luciferase reporter gene driven by five repeats of the hypoxia responsive element or the 5XHRE-luciferase reporter (Fig. 8B). However, no significant changes in myogenic differentiation were found when myoblasts were infected with retrovirus containing the gene coding for either constitutively active form of HIF-1α (Fig. 8A). These experiments indicate that HIF-1 function is not critically involved in hypoxic modulation of myogenesis. To further investigate the role of the HIF-1 pathway in myogenesis, we ectopically expressed DEC1/Stra13, a target gene of HIF-1, in C2C12 cells. DEC1/Stra13 is a member of the basic helix-loop-helix proteins that share homologies with the hairy and enhancer-of-split (HES) family of transcription repressors (7, 49). In our previous study, we have found that DEC1/Stra13 functions as an effector of HIF-1 to repress adipogenesis (55). However, ectopic expression of DEC1/Stra13 does not prevent C2C12 cells from forming myofibers (Fig. 8C). Furthermore, we examined the expression of DEC2, a homolog of DEC1/Stra13 that can suppress the transcription activity of myoD protein (2). As shown by RT-PCR, the expression of DEC2 does not change significantly in C2C12 cells up to 3 days at 0.5 or 0.01% O₂ (Fig. 9A). Together these observations indicate that the HIF-1 pathway does not play an important role in the hypoxic inhibition of myogenesis.

FIG. 8.

Effects of constitutively active HIF-1α mutants on myogenesis. (A) C2C12 cells were infected with retrovirus containing the indicated HIF-1α construct and allowed to differentiate at different pO₂ levels for 2 days. Cells were fixed and stained by immunofluorescence for MHC (green) with propidium iodide (red) as a nuclear counterstain. Results are representative of three independent experiments. (B) The relative transcription activity of the HIF-1α constructs was assessed by luciferase assay with 5XHRE-luciferase as a reporter. The transcription activity of the endogenous HIF-1 was reflected in the vector control samples. Normoxia, 21% O₂; hypoxia, 0.01% O₂. (C) C2C12 cells were infected with retroviral constructs containing the full-length DEC1/Stra13 and induced to differentiate for 3 days at 21% O₂. Bar, 50 μm.

FIG. 9.

Evaluation of the Notch signaling pathway and HES family proteins in hypoxic inhibition of myogenesis. (A) Total RNA was collected from differentiating C2C12 cells under the indicated conditions. RT-PCR was performed with gene-specific primers. (B) Whole-cell lysates were prepared from differentiating C2C12 cells under the indicated conditions. The NICD was immunoblotted with an anti-Notch1 monoclonal antibody (clone mN1A; Becton Dickinson). (C) C2C12 cells were allowed to differentiate for 3 days at 21% or 0.5% O₂ in the presence of DAPT to block notch activation. Myofiber formation was analyzed by anti-MHC immunofluorescence (green).

The HES family proteins inhibit myogenesis by functioning as effectors of the Notch signaling pathway (27, 35). Of particular interest is the negative regulator of myogenic differentiation, HERP2, also known as Hesr1, Hey1, HRT1, and CHF2 (27, 52). HERP2 attenuates the transcription activity of myoD (52). Interestingly, HERP2 also interacts with HIF-1β in the yeast two-hybrid assay (16). In order to investigate whether the Notch signaling pathway is involved in hypoxic regulation of myogenesis, we therefore examined the effects of hypoxia on the expression of notch1, notch2, notch3, and HERP2 by RT-PCR with β-actin as a control. Under hypoxia conditions (0.5 and 0.01% O₂), there were insignificant changes in the expression of HERP2 and notch3, compared to the levels in the normoxia (21% O₂) control (Fig. 9A). In contrast, severe hypoxia (0.01% O₂) significantly decreased the expression of notch1 and notch2. These results indicate that the Notch signaling pathway is unlikely to be involved in the repression of myogenesis by hypoxia. Consistent with this notion, we found little change at the level of the Notch1 intracellular cleaved domain (NICD) at 1 to 2% O₂, but robust reduction at 0.01% O₂ (Fig. 9B). Furthermore, we treated C2C12 cells with DAPT, a γ-secretase inhibitor, to block Notch signaling by preventing the formation of NICD (43). DAPT had no significant effects on myogenesis either at 21% or 0.5% O₂ (Fig. 9C). Taken together, these results support the conclusion that hypoxic inhibition of myoD and E2A is independent of Notch signaling.

Hypoxia attenuates the transcription activity of myoD promoter.

The myoD protein has a short half-life (38), and hypoxia does not significantly change its stability (Fig. 5C). The overall rate of protein synthesis in C2C12 myoblasts is also not significantly affected by hypoxia (1). Furthermore, we found that inhibition of protein degradation by the protease inhibitor MG132 failed to overcome the transient repression of myoD protein expression (Fig. 5D) or to restore myogenic differentiation at 0.5% O₂ (unpublished observation). The reduction of myoD protein can thus be attributed to decreased myoD mRNA. We investigated the effects of hypoxia on myoD promoter activity, using a luciferase construct containing a 4-kb distal enhancer element and a 2.5-kb proximal promoter sequence (26). In order to better reflect the effects of hypoxia on the endogenous myoD promoter during myogenic differentiation, we transfected C2C12 cells with linearized luciferase reporter gene plasmids and established cell lines that stably express the myoD promoter-driven luciferase reporter constructs. We normalized the hypoxia/normoxia ratio for each myoD promoter construct to the hypoxia/normoxia ratio for SV40 promoter-driven luciferase (pGL2) on days 1 and 4 of myogenic differentiation, respectively. We adopted this type of data analysis to more properly represent the specific effects of hypoxia on the myoD promoter, using a non-myogenically related promoter, SV40, as the control. The full-length (enhancer plus proximal) promoter was repressed approximately 50% on day 1 at 0.5% O₂ (Fig. 10A). Nonetheless, the transcriptional activity of the full-length promoter recovered to the level of the 21% O₂ control (Fig. 10A), thus recapitulating the adaptive expression of myoD mRNA under hypoxia (Fig. 4A). The proximal sequence was not affected significantly by hypoxia. However, the transcriptional activity of the distal enhancer sequence was suppressed by about 40% on day 1 and also recovered on day 4 at 0.5% O₂ (Fig. 10A). Taken together, these results suggest that the distal 4-kb enhancer fragment is susceptible to regulation by hypoxia.

FIG. 10.

Effects of hypoxia on transcription activity of the myoD promoter. C2C12 cells that stably express the indicated myoD promoter-driven luciferase reporter gene were allowed to undergo myogenic differentiation at 21 or 0.5% O₂. Luciferase assays were performed at the indicated time. (A) To better reflect the effects of hypoxia on myoD promoter activity, the hypoxia/normoxia ratio for each of the myoD promoter constructs was normalized to the hypoxia/normoxia ratio for SV40 promoter-driven luciferase (pGL2). (B) C2C12 cells were induced to differentiate under the indicated conditions. Total cell lysates were prepared in SDS sample buffer and subjected to polyacrylamide gel electrophoresis and Western blotting for acetylated H3 and H4. Actin served as a loading control. Cells treated with 5 mM sodium butyrate (NaBut) served as a positive control for acetylated histones, with 1/5 of the amount loaded in the other lanes. (C) C2C12 cells were induced to differentiate under the indicated conditions. ChIP assays were performed with the rabbit anti-acetylated H3 or control rabbit IgG. The myoD promoter was amplified by PCR with nested primers. The band intensities were analyzed with NIH Image 1.63 from two independent gel analyses (D).

The adaptive response of myoD mRNA expression under hypoxia suggests a transient regulation at the level of chromatin. We examined the effects of hypoxia on the acetylation of histones H3 and H4. While the levels of acetylated H3 and acetylated H4 did not change significantly at 0.5% O₂, compared to those at 21% O₂ (Fig. 10B), we found reduced amounts of myoD promoter gene associated with the acetylated H3 at 0.5% O₂ by ChIP assay (Fig. 10C and D). As acetylated histones are often associated with actively transcribed genes, these data suggest that hypoxia transiently reduces histone acetylation of the myoD promoter.

DISCUSSION

Myogenic differentiation of myoblasts is transiently repressed by hypoxia, indicative of an intrinsic adaptation mechanism. This adaptive mechanism is accompanied by the initial reduction of the myoD, E2A, and myogenin genes, followed by the resumption of their expression under hypoxia. Understanding of this unique phenomenon may hold the key to unraveling the complex nature of ischemic muscle damage and regeneration.

Muscle possesses remarkable plasticity and is able to repair ischemic damage via muscle regeneration. However, regenerated skeletal muscle fibers in patients with chronic lower limb ischemia are smaller than normal myofibers (40). Ischemia studies in rodents have also shown decreased cross-sectional areas of newly developed muscle fibers (44, 46). A significant decrease in limb muscle and cross-sectional myofiber areas has been found in human subjects undergoing simulated mountain climbing (33) and in mountaineers (25). Although the exact mechanisms are not clear, hypoxia is likely to be a critical factor for the characteristics of the regenerated muscle fibers. As this study shows, myofibers developed under hypoxia tend to have smaller cross-sections and to be poorly organized. Such observation suggests that hypoxia can potentially affect expression and function of yet unidentified genes involved in myocyte fusion and myofiber growth.

The ability of myoblasts to differentiate under hypoxia results from the active adaptation of myogenesis to hypoxic stress. For the high-altitude dwellers such as Tibetans and Quechuas who have been living at ≥3,500 m for thousands of generations, the capillary density in skeletal muscle is within the normal range and the mitochondrial content of skeletal muscle is also close to that of the lowland population, yet they develop muscle fibers with smaller cross-sectional areas (29). Prolonged exposure to high-altitude hypoxia (24) or simulated hypoxia in a chamber (33) can also cause a reduction of muscle fiber size and loss of skeletal muscle mass, despite a decrease in the perfusion area per capillary (24). These observations strongly suggest that the decrease in skeletal muscle mass and myofiber size directly results from hypoxia. The structural and functional features of skeletal muscle presented by Tibetans and Quechuas are consequences of the adaptive myogenic differentiation under chronic hypoxia. Other studies have found that hypoxia seems to favor the expression of MHC IIA over MHC I in the soleus muscle of young rats exposed to high-altitude hypoxia (barometric pressure = 370 mm Hg; pO₂ = 79 mm Hg) for 4 weeks (5). Taken together, these findings underscore the importance of microenvironment in the regulation of gene functions and cellular commitment to differentiation.

Muscle satellite cells are believed to be the precursors or stem cells for muscle regeneration (23). Early stem cells give rise to intermediate stem cells called myoblasts as a result of myoD and myf-5 expression. The regulation of myoD activity is the key event for the terminal differentiation of myoblasts to myofibers. The transcription activity of myoD is regulated by posttranslational modifications and by other proteins that interact with myoD (38). In response to myogenic stimulation, myoD protein becomes activated, dimerizes with one of the E2A proteins (E12/E47), and induces the expression of the myogenin gene, which then drives the terminal differentiation of myoblasts. Under hypoxia, rapid reduction occurs at both the myoD mRNA and myoD protein levels. The myoD protein has a short half-life (38). Our data indicate that hypoxia does not significantly alter its stability either in C2C12 or in primary myoblasts. Using a different approach, Di Carlo et al. have reported a slight decrease in myoD stability under hypoxia (18). We have further found that hypoxia does not significantly change the stability of myoD mRNA or the apparent translation of myoD protein. The reduction of myoD protein can thus be attributed to decreased myoD mRNA. As our data have shown, hypoxia attenuates the transcription activity of the myoD promoter gene (4-kb distal enhancer plus 2.5-kb proximal promoter) with 24 h. Furthermore, its transcriptional activity recovers to the normoxic control level by day 4 at 0.5% O₂, consistent with the kinetics of myoD mRNA expression under hypoxia. As our data have further shown, the extent of myoD promoter associated with acetylated H3 is reduced by hypoxia, suggesting that the transient hypoxia repression of myoD gene expression involves transient deacetylation of histones. However, the mechanism underlying the effects of hypoxia on deacetylation of histones specifically associated with the myoD promoter seems to be promoter specific, because hypoxia does not affect the global acetylation status of histones. We also investigated the role of histone acetylation in myoD transcription under hypoxia, using a histone deacetylase inhibitor, trichostatin A. We found that trichostatin A inhibited myogenesis at 21 and 0.5% O₂ (unpublished observation), indicating that orchestrated gene repression as well as activation is needed for myogenesis.

The transcriptional control of myoD expression is complex. Elegant genetic studies by the Goldhamer group have identified the 24-kb 5′ genomic sequence that recapitulates myoD expression in transgenic mouse embryos (15). The enhancer region located 20 kb upstream of the myoD coding sequence seems to be required for initiation of myoD expression in myogenic precursor cells (15). However, genetic deletion of the core enhancer element only causes delayed expression of myoD and belated myogenic differentiation (13). Furthermore, the cis elements in this 24-kb genomic fragment upstream of myoD coding region still cannot explain the regulation of myoD transcription by genes that affect muscle development, such as pax-3 and myf-5 (13). It also remains unclear how myoD transcription is affected by other myogenic regulators, such as the wnt family proteins, sonic hedgehog (shh), and bone morphogenic protein 4 (BMP4) (14). The complex regulation of myoD transcription may be, at least in part, correlated with epigenetic modification of the myoD enhancer/promoter region because the myoD enhancer does not have muscle specificity in tissue culture, despite its specific activity in muscle lineages during embryonic development (14). Consistent with these past studies, our findings demonstrate that hypoxia induces transient histone deacetylation in the myoD promoter region and repression of myoD transcription.

As our data indicate, ectopic expression of HIF-1α constructs does not have significant effects on myoblast differentiation under either hypoxia or normoxia. However, HIF-1 may play a role in muscle development during early embryogenesis. Before their demise by embryonic day 11.0 (E11.0) due to massive developmental abnormalities, HIF1a^−/− embryos have much fewer somites and exhibit morphological abnormalities of pericardium between E9.75 and E10.0 compared to HIF1a^+/+ embryos (28). However, our data indicate that HIF-1 is not directly involved in postnatal myogenesis. Consistent with our findings, Mason et al. have found that targeted deletion of HIF-1α in skeletal muscle affects the energy metabolism of muscle cells more profoundly, while the development of skeletal muscle seems to be quite comparable to that of the wild-type control (34).

Unlike hypoxic inhibition of adipogenesis that is mediated by repression of PPARγ2 by the HIF-1 target gene DEC1/Stra13, a member of the HES family (55), hypoxic repression of myoD seems to be independent of the HIF-1 pathway. There are other notable differences between adipogenesis and myogenesis under hypoxia. Full inhibition of adipogenesis occurs at 2% O₂, and there is no recovery of PPARγ2 expression under any levels of hypoxia (55). Inhibition of myogenesis is proportional to the severity of hypoxia, and myoD expression can recover efficiently along with myogenesis, even at 0.5% O₂. This observation suggests that myoblast cells have acquired different abilities for environmental sensing and adaptation despite sharing the same mesodermal origin with preadipocytes.

Myoblasts are highly unique in that they can adapt to limiting oxygen conditions and retain their differentiation potentials. The mechanisms for the hypoxic regulation of myogenesis are extremely complex. Multiple pathways may be operative sequentially from hypoxic repression to adaptation. Nevertheless, our studies present a molecular basis for the clinical observations in patients suffering from chronic tissue hypoxia, as well as the effects of high altitude on muscle fiber development. Further studies are needed to delineate the intricate modes of myogenic responses to hypoxia at both the myoblast and the myofiber levels. Nevertheless, our data reveal that highly committed myoblasts with high levels of both myoD and myogenin gene expression exhibit a strong propensity for differentiation under hypoxia. This finding suggests that targeted expression of MRFs might be a viable gene therapy approach for the repair and regeneration of ischemically damaged muscle.