Hypoxia-inducible Factor 1-α Induces miR-210 in Normoxic Differentiating Myoblasts^*

Immunoblotting

Cells were lysed in 2× Laemmli buffer and boiled for 5 min. Equal amounts of proteins were separated by SDS-PAGE and transferred to nitrocellulose by standard procedures. Proteins of interest were detected with the following specific antibodies: anti-Hif1a (H1α6, Novus Biological), and anti-α-tubulin (Ab-1, Oncogene Research Product).

Indirect Immunofluorescence

Cells were fixed with 4% paraformaldehyde, and immunofluorescence staining was performed as described previously (32). Myotubes were identified using a myosin heavy chain, Myh2 antibody (MF 20 mouse hybridoma). Cells were counterstained with Hoechst 33342 (Sigma-Aldrich). Images were acquired by a fluorescence microscope (Axioplan 2, Carl Zeiss), analyzed by IAS software (Delta System). Myotubes and nuclei were counted by two blinded readers obtaining similar results. The fusion index was calculated as the percentage of nuclei belonging to myosin heavy chain-positive cells with three or more nuclei (33).

Oxidative Stress

Cell cultures were incubated at 37 °C with 2 μm dihydroethidium (DHE) (Sigma-Aldrich) or 5 μm MitoSOX (Molecular Probes, Invitrogen) for 15 min and 10 min, respectively. Then, cells were fixed with 4% paraformaldehyde, nuclei were stained with Hoechst 33342, and fluorescence was revealed by fluorescence microscopy using the same exposure conditions for each sample and quantified using Scion Image software as described previously (7). DHE and MitoSOX fluorescence was normalized for the number of Hoechst 33342-positive nuclei.

miRNA Down-modulation and RNAi

Locked nucleic acid (LNA) oligonucleotides against miR-210 (in vivo LNA miR-210 inhibitor) or a control scramble LNA sequence (40 nm, both from Exiqon) were transfected using siRNA transfection reagent (SC 29528, Santa Cruz Biotechnology) in C2C12 in GM. After 16 h, cells were washed, and fresh medium was added, and experiments were performed 24 h later.

For siRNA silencing, C2C12 cells were transfected with the HiPerfect transfection reagent (Qiagen) complexed with siRNAs according to the instructions of the manufacturer. Briefly, in C2C12 RNAi for a single molecule was performed by complexing 6 μl of HiPerfect and 25 nm siRNA targeting Hif1a (SC 35562, Santa Cruz Biotechnology) or a scramble sequence per 35-mm dish. After 16 h, cells were washed, and fresh medium was added.

For shRNA-mediated knockdown, the following shRNA pLKO.1 plasmids were employed: pLKO.1 empty (Addgene 8543), pLKO.1 scrambled shRNA (Addgene 1864), and pLKO.1 Hif1a shRNA (Sigma-Aldrich TRCN0000232220, TRCN0000232222, and TRCN0000232223). Lentiviral generation and infection were performed as described previously (34). Uninfected cells were selected out by puromycin selection.

miRNA and mRNA Quantification

Total RNA was extracted using TRIzol (Invitrogen). miRNA levels were analyzed using the Applied Biosystems TaqMan quantitative real-time PCR method (qPCR, 1 ng/assay) performed according to the instructions of the manufacturer and quantified with the ABI Prism 7000 SDS (Applied Biosystems). Mature miRNA levels were normalized to miR-16, whereas pri-miR-210 levels were normalized to GAPDH.

Hif1a mRNA Levels Were Analyzed Using Applied Biosystems TaqMan Quantitative Real-time PCR Method (1 ng/assay)

mRNAs levels were analyzed using the SYBR-Green qPCR method (5 ng/assay, Qiagen) performed according to the instructions of the manufacturer using supplemental Table S1 primers and quantified with ABI Prism 7000 SDS (Applied Biosystems). mRNA expression was normalized for β-2-microglobulin (B2m) or Rpl13 levels. For both miRNAs and mRNAs, relative expression was calculated using the comparative Ct method (2^−−ΔΔCt).

ChIP Assay

The procedure for ChIP was performed as described previously (35). Briefly, 1% formaldehyde was added directly to the cells and incubated at 22 °C for 10 min. The reaction was stopped adding 0.125 m glycine. Then, the cells were rinsed with cold 1× PBS, incubated with 0.2× trypsin-EDTA in 1× PBS, and scraped. cells were centrifuged, washed in cold 1× PBS plus 0.5 mm PMSF and resuspended in lysis buffer (5 mm piperazine N,N bis zethone sulfonic acid (pH 8.85) mm KCl, 0.5% Nonidet P-40). Next, nuclei were solicited in the sonication buffer (0.1% SDS, 10 mm EDTA, 50 mm Tris-HCl (pH 8), 0.5% deoxycholic acid) for 10 min by using a microultrasonic cell disruptor. The chromatin was sheared to an average size of 500 base pairs, and immunoprecipitation was performed with protein G-agarose (KPL). The chromatin solution was precleared by adding protein G for 1 h at 4 °C and incubated at 4 °C overnight with 4 μg of Hif1a antibody (H1α6, Novus Biological) or non-specific immunoglobulins (IgGs, Santa Cruz Biotechnology) as negative control. Input was collected from a control sample supernatant (not immunoprecipitated antibody). Immunoprecipitates were recovered by incubation for 2 h at 4 °C with protein G-agarose precleared previously in immunoprecipitation buffer (1 μg/μl bovine serum albumin, 1 μg/μl salmon testis DNA, protease inhibitors, and PMSF). Reversal of formaldehyde cross-linking, RNase A, and proteinase K treatments were performed as described previously (35). DNA was phenol-extracted, ethanol-precipitated, and analyzed by PCR. DNA representing 0.005–0.01% of the total chromatin sample (input) or 1–10% of the immunoprecipitates was amplified using specific primers indicated in supplemental Table S2.

Reporter Assay

miR-210 promoter constructs were generated using pGL2 firefly luciferase vector (Promega) adopting standard techniques. Relevant mouse genomic DNA fragments were amplified using supplemental Table S3 primers and cloned between the BglII and HindIII sites. For the HRE mutational analysis, HRE 1 and 2 were mutated within the D construct promoter sequence either alone (D mut 1 and D mut 2) or in combination (D mut 1 + 2). The HREs (A/GCGTG) were mutated to A/GAAAG. The corresponding DNA fragments were obtained from the Integrated DNA Technologies (IDT) gene synthesis service (TEMA ricerca) and then subcloned in the pGL2 luciferase plasmid. pGL2 luciferase plasmids were cotransfected with a pRL-null plasmid encoding Renilla luciferase (Promega). All transfections were carried out using FuGENE 6 transfection reagent (Roche), and luciferase activity was measured using the dual luciferase assay system (Promega) according to the instructions of the manufacturers.

Analysis of miR-210 CpG Island Methylation

To identify CpG islands in the 2.373 kb upstream mature miR-210, the DNA sequence was submitted to the CpG Island Searcher using the following parameters: length, 300 bp; ObsCpG/ExpCpG, 70%; and %GC, 60%. C2C12 DNA bisulfite conversion was performed directly using the EZ DNA Methylation-Direct kit following the recommendations of the manufacturer (Zymo Research). Products of the bisulfite reactions were PCR-amplified using the HotMaster TaqDNA polymerase kit (5 PRIME) using supplemental Table S4 primers. The products were resolved on a 1.6% agarose gel, purified using a MinElute gel extraction kit (Qiagen), subcloned into the pDRIVE vector (using a Qiagen PCR cloning kit), and sequenced.

Mouse Model of Skeletal Muscle Regeneration

All experimental procedures complied with the guidelines of the Italian National Institutes of Health and with the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Academy of Sciences, Bethesda, MD) and were approved by the institutional Animal Care and Use Committee. Two-month-old CD1 male mice were used. In vivo down-modulation of miR-210 was carried out by tail vein injection of 12 mg/kg “in vivo LNA miR-210 inhibitor” (Exiqon) or an LNA scrambled sequence (SCR). For longer time courses, a second LNA oligonucleotide dose was administered at day 7. Two days after the first LNA administration, tibialis anterior muscle injury was induced by cardiotoxin (CTX, Calbiochem) injection as described previously (32). Next, both treated and controlateral muscles were harvested and split into two parts. Half was fixed, paraffin-embedded, and used for histological analysis. The other half was used for RNA extraction. To measure CDX-induced damage, myofiber permeability was measured by intraperitoneal injection of Evans Blue dye (Sigma-Aldrich) 16 h before sacrifice, as described previously (36). Thereafter, muscles were harvested and split into two parts. Half was frozen in OCT embedding medium, and half was used for RNA extraction.

Histology and Morphometric Analysis

H&E sections of paraffin-embedded muscles were prepared as described previously (32), and regenerating centrally nucleated myofibers were counted. Frozen sections of Evans Blue dye-stained muscles were fixed in cold acetone at −20 °C for 10 min, washed in PBS, and mounted with fluorescence mounting medium. A Zeiss Axioplan 2 fluorescence microscope with image analyzer KS300 software was used to acquire images and to measure areas. All histological and morphometric analyses were carried out by two blinded readers with comparable results.

miR-210 Induction during Differentiation Occurs at the Transcriptional Level and Is Hif1a-dependent

We sought to determine how miR-210 is regulated during differentiation. Fig. 2A shows that pri-miR-210 was induced upon myogenic differentiation to levels similar to these of the corresponding mature miRNA, indicating a transcriptional regulation of miR-210 during differentiation. These data were also confirmed with an independent primer couple (not shown). Next, given the importance of HIF1A for both miR-210 induction following hypoxia (1) and skeletal muscle physiopathology (37–40), we tested whether Hif1a silencing affected miR-210 induction in differentiating C2C12 cells.

Open in a separate window

FIGURE 2.

miR-210 transcriptional activation upon myogenic differentiation is Hif1a-dependent. C2C12 cells were cultured in GM and then switched to DM for the indicated time. A, RNA was extracted, and pri-miR-210 levels were measured by qPCR. Pri-miR-210 was induced during differentiation (n = 5; ** p < 0.01; *** p < 0.001). B, alternatively, protein extracts were derived, and Hif1a protein levels were measured by Western blot analysis. α-tubulin levels were used for gel loading control. Hif1a protein increased during C2C12 differentiation. C, C2C12 cells were transfected with a Hif1a-specific siRNA (Hif1a siRNA) or with a control scramble sequence (SCR). Then, cells were allowed to differentiate in DM for the indicated time, and miR-210 levels were measured. Hif1a silencing significantly decreased miR-210 levels (n = 9; ** p < 0.01).

Western blot analysis showed that Hif1a protein increased upon C2C12 differentiation (Fig. 2B). Next we assayed the effect of Hif1a silencing on miR-210 levels in differentiating myoblasts. The Hif1a transcript was not modulated during C2C12 differentiation and was significantly knocked down following specific siRNA transfection (supplemental Fig. S2). When miR-210 expression was measured, we found that miR-210 levels were decreased significantly at both the 24 and 48 h differentiation time points in Hif1a-silenced cells (Fig. 2C). miR-210 expression was decreased in myoblasts as well, but the difference did not reach statistical significance. Decreased miR-210 levels in DM were not simply due to an impairment of the differentiation process because the expression of myogenic markers such as miR-1, myogenin, and myosin heavy chain was unaffected (supplemental Fig. S2).

To confirm these findings, Hif1a was also knocked down in transducing C2C12 cells with a lentiviral vector expressing a specific shRNA (supplemental Fig. S3A). The adopted vector also carried a puromycin resistance gene, allowing the selection of infected cells. As expected, Hif1a silencing decreased miR-210 levels in hypoxic myoblasts (supplemental Fig. S3B). When normoxic differentiating myoblasts were measured, we found that miR-210 levels were decreased significantly by Hif1a shRNA both in GM and in DM. Indeed, after 48 h of differentiation, miR-210 levels were still lower than these observed in myoblasts expressing a scrambled control (supplemental Fig. S3C). This stronger inhibition of miR-210 expression was likely due to the higher efficiency of transduction yielded by infection followed by puromycin selection compared with transient transfection. Again, unaffected miR-1 induction upon differentiation indicated that myogenic differentiation was not grossly inhibited by Hif1a knockdown (supplemental Fig. S3D). These findings were further confirmed by two independent shRNA-targeting Hif1a (not shown).

Hif1a Binds to miR-210 Promoter, Activating Its Expression in Differentiating Myoblasts

Analysis of the ≈2.4 kbase promoter sequence upstream of the mature miR-210 sequence identified four potential HRE sites. To elucidate whether these sites were functionally relevant, a chromatin immunoprecipitation experiment was performed. As shown in Fig. 3, Hif1a antibody (but not control IgGs) immunoprecipitated the miR-210 promoter fragments in hypoxic and in differentiated C2C12 cells but very little in the normoxic undifferentiated controls. Specifically, two primer couples were used, encompassing the two proximal HRE (HRE1 and 2) and the two distal sites (HRE3 and 4). A more defined analysis was precluded by the close proximity of the HREs. We found that, although Hif1a bound to both fragments upon hypoxia, in differentiated myotubes Hif1a preferentially bound to the proximal fragment.

Open in a separate window

FIGURE 3.

ChIP analysis of the miR-210 promoter. Upper panel, schematic representation of the HREs located in the ≈2.4-kbase promoter region upstream the mature miR-210. HRE 1, 2, 3, and 4 are located 437 bp, 495 bp, 841 bp, and 905 bp upstream of the mature miR-210, respectively. PCR was performed with primers specific to amplify promoter fragments containing the proximal two HREs (1 and 2) or the more distal HREs (3 and 4). The genomic sequence corresponding to the amplified fragments is displayed. The lower panel shows the Hif1a ChIP results. A ChIP analysis was performed using C2C12 cells cultured under GM or DM in normoxia for 48 h or GM and 1% oxygen tension for 24 h. IgG refers to samples derived for the IP negative control. In differentiated normoxic myotubes, Hif1a specifically bound only to the proximal fragment (HRE 1 and 2), whereas under hypoxia, Hif1a bound to both fragments.

To better identify and characterize the functional HRE, we amplified ≈2.4 kbases of the genomic regions upstream of miR-210 sequence and generated serial deletion constructs (Fig. 4A). The promoter fragments were cloned into a luciferase vector, and the effect of the deletions on reporter gene activity during C2C12 myogenic differentiation was analyzed. Luciferase assays showed that the miR-210 promoter was induced in DM and that a 565-base fragment (D construct) that harbors the two proximal HRE sites (HRE1 and 2) was essential for sustained luciferase activity in differentiated myotubes (Fig. 4A). We also found that miR-210 D construct activity was decreased significantly in C2C12 myotubes where Hif1a was silenced, further corroborating the Hif1a role in miR-210 promoter modulation (Fig. 4B). Further deletions encompassing HRE2 (construct E) significantly decreased miR-210 promoter activity, although some residual functionality and inducibility were present (Fig. 4A). Interestingly, a similar pattern was observed when Hif1a overexpressing myoblasts were assayed, indicating that HRE1 was not functional (supplemental Fig. S4A). To confirm these findings, HRE1 and HRE2 were mutated, either alone or in combination. Fig. 4C shows that HRE2 mutation decreased the activity of the D construct, whereas HRE1 mutation was ineffective. Again, similar results were obtained when hypoxic myoblasts were assayed, although promoter activity was much stronger (supplemental Fig. S4B).

Open in a separate window

FIGURE 4.

Hif1a positively modulates miR-210 promoter activity during differentiation. A, miR-210 promoter fragments were cloned into the pGL2-luciferase vector. Serial deletion constructs (identified by the letters A–F) are schematically represented above the bar graph. Predicted HREs are shown. Lower panel, the relative promoter activities in GM and DM 24 h are displayed. The empty pGL2-basic vector was used to monitor basal activity of the luciferase reporter gene (Basic). Luciferase activity is expressed as firefly/Renilla luciferase ratio. Constructs A–D displayed similar expression levels in DM, and activity dropped significantly in the E and F constructs (n = 4; **p < 0.01). B, C2C12 cells infected with lentiviral vectors expressing either a Hif1a-specific shRNA (Hif1a shRNA) or a scramble (SCR) sequence were transfected with construct D (miR-210 565-bp promoter), and luciferase assays were performed in GM and in DM both at 24 and 48 h. D construct activity was significantly decreased in C2C12 myotubes when Hif1a was silenced (n = 3; ** p < 0.01; *** p < 0.001). C, upper panel, the two proximal HREs were mutated either alone (D mut 1 and D mut 2) or in combination (D mut 1+2). Lower panel, promoter activity was detected by luciferase assays in C2C12 cells in GM and DM after 24 h. HRE2 mutation significantly decreased D construct activity, whereas HRE1 mutation was ineffective (n = 6; *, p < 0.05; **, p < 0.01; ns, not significant).

Finally, promoter analysis indicated the presence of a potential Myod binding site 523 nucleotides upstream of miR-210. However, mutational analysis showed that this sequence did not affect miR-210 promoter expression either in GM or in DM (supplemental Fig. S4C and Table S3).

Bioinformatic analysis of the ≈2.4 kbases upstream of the miR-210 mature sequence also indicated a potential CpG island from base −548 to base −280 (supplemental Fig. S4D), suggesting that DNA methylation may contribute to miR-210 promoter regulation during muscle differentiation. Indeed, a CpG island has been identified upstream of human miR-210 as well (41). When the cytosine methylation of this CpG-rich region was analyzed using the bisulfite technique, we found that methylation was almost undetectable in myoblasts and that the methylation status did not change during differentiation (not shown). In keeping with these findings, C2C12 treatment with 5-azacytidine demethylating drug for 24–72 h did not affect miR-210 expression (not shown). Thus, cytosine methylation does not seem to play a major role in miR-210 regulation during myogenic differentiation.

miR-210 Blockade Does Not Affect Myogenic Differentiation

To assess the functional role of miR-210 in skeletal muscle differentiation, we inhibited miR-210 in C2C12 cells using a LNA-modified antisense oligonucleotide (anti-miR-210). Although miR-210 levels were decreased more than 10-fold in differentiated cells (supplemental Fig. S5), we did not observe any overt alteration in myotube morphology (Fig. 5A, left panel). In keeping with this observation, the expression levels of myogenesis markers such as myogenin, desmin, myod, miR-1, miR-133a, and miR-133b were unaffected by miR-210 blockade (supplemental Fig. S5). Also, the myotube number (not shown), fusion index, and average number of nuclei per myotube after 48 h of differentiation did not change significantly (supplemental Fig. S6).

Open in a separate window

FIGURE 5.

miR-210 increases myotube survival under chemical anoxia induced by sodium azide treatment. C2C12 myoblasts transfected with anti-miR-210 or a LNA control scramble sequence (SCR) were cultured for 24 h in DM and then treated with 40 mm sodium azide (NaN₃) for the indicated time to induce chemical anoxia. Next, cells were fixed and stained with MHC antibody, and the number of myotubes was counted. A, representative MHC immunofluorescence. Scale bar = 30 μm. B, bar graph showing a more pronounced NaN₃ toxicity in anti-miR-210 treated cultures (n = 3; * p < 0.05; #, p < 0.02).

We also assayed whether miR-210 blockade in vivo affected skeletal muscle regeneration. miR-210 was inhibited by systemic anti-miR-210 administration, decreasing miR-210 levels > 10-fold (supplemental Fig. S7A). Skeletal muscle injury was induced by CTX injection, and Evans Blue staining showed that similar damage was induced in the two groups (supplemental Fig. S7B). Skeletal muscle regeneration was assayed 7 and 14 days after CTX injection. Although miR-210 was decreased efficiently, the number of centrally nucleated regenerating myofibers was similar in anti-miR-210 and LNA-scrambled groups (supplemental Fig. S7C). In keeping with this finding, regeneration markers such as miR-206 and myogenin were induced similarly in the two groups both at days 7 and 14 (supplemental Fig. S7D). In conclusion, at least in the adopted experimental conditions, we found no evidence that miR-210 is necessary for myogenic differentiation.