Separating myoblast differentiation from muscle cell fusion using IGF-I and the p38 MAP kinase inhibitor SB202190

Abstract

The p38 MAP kinases play critical roles in skeletal muscle biology, but the specific processes regulated by these kinases remain poorly defined. Here we find that activity of p38α/β is important not only in early phases of myoblast differentiation, but also in later stages of myocyte fusion and myofibrillogenesis. By treatment of C2 myoblasts with the promyogenic growth factor insulin-like growth factor (IGF)-I, the early block in differentiation imposed by the p38 chemical inhibitor SB202190 could be overcome. Yet, under these conditions, IGF-I could not prevent the later impairment of muscle cell fusion, as marked by the nearly complete absence of multinucleated myofibers. Removal of SB202190 from the medium of differentiating myoblasts reversed the fusion block, as multinucleated myofibers were detected several hours later and reached ∼90% of the culture within 30 h. Analysis by quantitative mass spectroscopy of proteins that changed in abundance following removal of the inhibitor revealed a cohort of upregulated muscle-enriched molecules that may be important for both myofibrillogenesis and fusion. We have thus developed a model system that allows separation of myoblast differentiation from muscle cell fusion and should be useful in identifying specific steps regulated by p38 MAP kinase-mediated signaling in myogenesis.

muscle growth and repair rely on stem cells termed satellite cells (50), which, upon activation, undergo sequential proliferation, terminal differentiation, and fusion with existing myofibers (3, 50). Satellite cell activity is controlled by a complex milieu of potentially competing signals mediated by cell-cell contact, growth factors, and hormones (3) that interact with genetic programs directed by myogenic transcription factors of the MyoD family (14, 47). Among secreted molecules with promyogenic activity are the insulin-like growth factors, IGF-I and IGF-II. These proteins play key roles in muscle development in the fetus (29, 45) and are important in coordinating muscle repair and reinnervation after injury (2, 7, 37, 39).

IGF actions are mediated by the IGF-I receptor, a ligand-activated tyrosine protein kinase related to the insulin receptor (38). IGF binding triggers receptor tyrosine kinase function and initiates intracellular protein-protein interactions that activate multiple signaling pathways (38). Current evidence suggests that the phosphoinositide 3-kinase (PI3K)-Akt pathway is the critical IGF-activated signaling cascade in muscle (5, 16, 24, 40). Mice lacking Akt1 and Akt2 have a severe muscle deficiency phenotype that resembles loss of the IGF-I receptor (43); by contrast, transgenic mice expressing a constitutively active Akt in muscle have extensive hypertrophy (24).

The p38 MAP kinases are signal transducers that also promote muscle differentiation in vitro and influence muscle growth and repair in vivo (9, 25, 28, 44, 46, 55, 57, 58). Cell-cell contact can induce this pathway (23, 33) through stimulation of the upstream kinases Tak1 and Ask1 and MAP kinase kinase 3 (MKK3) and MKK6 (52). Two of four p38 family members, p38α and p38β, appear to be the most important in skeletal muscle (21, 31). Myoblasts from mice lacking p38α showed delayed growth and maturation, as well as reduced differentiation in culture (44). In contrast, forced expression of constitutively active MKK6 enhanced differentiation in vitro (46, 57), which was blocked by selective p38 chemical inhibitors (9, 28, 55, 57). Signaling by p38α has been shown to increase the activity of myogenic transcription factors (21, 30, 55–57, 59) and to facilitate chromatin opening on muscle gene promoters (48, 49). By contrast, p38γ has been found to block satellite cell differentiation (15, 25), indicating that individual p38s may mediate opposite biological effects in muscle.

Evidence for collaboration or cross talk between PI3K-Akt and p38-regulated signaling pathways in muscle is limited, consisting primarily of demonstrations that interference with one cascade can prevent differentiation mediated by the other (6, 17, 28). Here we show that IGF-mediated signaling can restore muscle cell differentiation that is otherwise blocked by a p38α/β MAP kinase inhibitor. Despite expressing muscle proteins and becoming both elongated and aligned, these cells were almost entirely mononucleated. Inhibitor removal caused rapid and extensive myocyte fusion and formation of multinucleated myofibers, and analysis by quantitative mass spectroscopy identified a number of upregulated muscle-enriched proteins of potential importance for both myofibrillogenesis and fusion. We have thus developed a model system that allows separation of myoblast differentiation from muscle cell fusion and should be useful in identifying specific steps regulated by p38 MAP kinase-mediated signaling in myogenesis.

MATERIALS AND METHODS

Materials.

Fetal and newborn calf serum was purchased from Hyclone (Logan, UT); horse serum, goat serum, Dulbecco's modified Eagle's medium (DMEM), PBS, trypsin-EDTA solution, SuperScript III reverse transcriptase (RT) kit, and SYBR Green Platinum qPCR mix from Invitrogen-Life Technologies (Carlsbad, CA); porcine gelatin and paraformaldehyde from Sigma-Aldrich (St. Louis, MO); and Hoechst 33258 nuclear dye from Polysciences (Warrington, PA). R3-IGF-I (GroPep, Adelaide, Australia) was solubilized in 10 mM HCl with 1 mg/ml bovine serum albumin, stored in aliquots at −80°C, and diluted into medium immediately prior to use. Okadaic acid was obtained from Alexis Biochemicals (San Diego, CA); protease inhibitor and NBT/BCIP tablets from Roche Applied Sciences (Indianapolis, IN); Qia-Quick PCR purification kit from Qiagen (Valencia, CA); BCA protein assay kit from Pierce Biotechnologies (Rockford, IL); tandem mass tag (TMT) reagent (TMT10plex) from Thermo Fisher Scientific (Waltham, MA); and restriction enzymes, buffers, ligases, and polymerases from New England Biolabs (Beverly, MA). Sep-Pak Light cartridges (Waters, Milford, MA) were used according to the supplier's instructions. Sequencing-grade modified trypsin (Promega, Madison, WI) was dissolved in 1 mM HCl at 1 μg/ml just prior to use. AquaBlock enzyme immunoassay/Western immunoblot solution was obtained from East Coast Biologicals (North Berwick, ME). SB202190 (Tocris Bioscience, Bristol, UK) was solubilized in DMSO and stored at 1,000× final concentration in aliquots at −80°C; it was diluted into medium immediately prior to use. Primary antibodies were purchased as follows: anti-myogenin (product no. F5D from W. E. Wright), anti-myosin heavy chain (product no. MF20 from D. A. Fischman), and anti-troponin-T (product no. CT3 from J. J.-C. Lin) from Developmental Studies Hybridoma Bank (Iowa City, IA); anti-Mef2c (catalog no. IA-14) from Covance (Madison, WI); anti-MyoD (catalog no. 5.8A) from BD-Pharmingen (San Jose, CA); anti-Akt (catalog no. 4691), anti-phosphorylated (p) Akt^Thr308 (catalog no. 2965), anti-p38 (catalog no. 8690), anti-p-p38 (catalog no. 4511), anti-Erk (catalog no. 4695), anti-p-Erk (catalog no. 4370), anti-mammalian target of rapamycin (mTor; catalog no. 2983), and anti-p-mTor (catalog no. 5536) from Cell Signaling (Beverly, MA); and anti-α-tubulin (catalog no. B-5-1-2) from Sigma-Aldrich. Secondary antibodies included goat anti-mouse IgGs conjugated with Alexa Fluor 488, Alexa Fluor 594, or Alexa Fluor 680 and goat anti-rabbit IgG conjugated with Alexa Fluor 680 (Invitrogen-Life Technologies) and IR800-conjugated goat anti-rabbit IgG (Rockland, Gilbertsville, PA). All other chemicals were reagent grade and were purchased from commercial suppliers.

Recombinant lentiviruses.

A recombinant lentivirus was prepared to express enhanced green fluorescent protein (EGFP) using the backbone plasmid pWPXLd (Addgene, Cambridge, MA). This lentivirus was generated in HEK 293FT cells (Life Technologies) that were co-transfected with pWPXLd and third-generation packaging plasmids (catalog nos. 12251, 12253, and 12259, Addgene), as described elsewhere (51). At 2 days after transfection, conditioned medium was collected, and lentiviruses were concentrated by centrifugation in a SW28 rotor at 19,000 rpm for 2 h at 4°C. The pellet was resuspended in PBS with 1% bovine serum albumin, and virus was stored in aliquots at −80°C (36). Prior to use, virus was diluted into DMEM plus 2% fetal calf serum and filtered through a 0.45-μm Gelman syringe filter.

Cell culture.

Cells were grown at 37°C in humidified air with 5% CO₂. C2 myoblasts were incubated on gelatin-coated tissue culture dishes in growth medium (DMEM with 10% heat-inactivated fetal calf serum and 10% newborn calf serum). C3H10T1/2 mouse embryonic fibroblasts (catalog no. CCL226, American Type Culture Collection) were grown on gelatin-coated tissue culture dishes in DMEM with 10% fetal calf serum and infected with the EGFP lentivirus, as described previously (13). Over 90% of cells expressed EGFP, which persisted at comparable levels for more than five additional passages. EGFP-expressing C3H10T1/2 cells were mixed with C2 myoblasts at a ratio of 1:20, thus providing a way to orient the population for live-cell imaging (see below). Muscle differentiation was induced when co-cultured cells reached ∼95% of confluent density by replacement of growth medium with differentiation medium (DM: DMEM + 2% horse serum) ± SB202190 (5 μM) ± R3-IGF-I (1 nM).

Immunocytochemistry.

Cells were fixed using 4% paraformaldehyde, permeabilized using methanol-acetone, and incubated with antibodies, as described elsewhere (12). Primary antibodies were added at a 1:1,000 dilution in goat serum containing blocking buffer for 16 h at 4°C. Secondary antibody was added at a dilution of 1:2,000, and Hoechst stain (1:1,000 dilution) was added in darkness for 1.5 h. Images were captured using a Nikon DS-Qi1Mc camera attached to a Nikon Eclipse Ti-U inverted microscope using NIS-Elements 3.1 software. Images were analyzed using a combination of NIS-Elements and Cell Profiler (www.cellprofiler.org). Results are presented as means ± SD of at least three experiments. The area occupied by differentiating muscle cells was calculated by measurement of the percentage of troponin-T-positive staining in a microscopic field.

Live cell imaging.

Cell cultures were monitored with the IncuCyte imaging system (Essen Bioscience, Ann Arbor, MI). Individual microscopic fields (4 per well) were imaged using a ×10 objective at 15-min intervals for up to 72 h. The resulting phase-contrast and fluorescence images were registered using EGFP fluorescence from labeled C3H10T1/2 cells. Individual cells (60 per experiment) were tracked using ImageJ (NIH, Bethesda, MD). Fusion was assessed by counting the fraction of cells in multinucleated myofibers at different times.

Analysis of gene expression.

Whole cell RNA was isolated as described previously (12). RNA integrity was established by agarose gel electrophoresis, and concentrations were determined by spectrophotometry at 260 nm. Total RNA (2 μg) was reverse-transcribed with oligo(dT) primers in a final volume of 20 μl using the SuperScript III RT kit (12). The cDNA (0.5 μl) was used as a template for conventional PCR, as described elsewhere (12), after pilot studies were performed to establish a cycle number for each primer set to achieve amplification in the linear range (∼22–27 cycles, depending on the primer). Primer pairs are shown in Table 1. Products were visualized after agarose gel electrophoresis and staining with ethidium bromide. Results are representative of at least three independent RT-PCR experiments with different isolates of RNA and cDNA.

Table 1.

Primers used for RT-PCR

	Primer
Gene	Forward	Reverse	Product, bp
MyoD	5′-TACAGTGGCGACTCAGATGC	5′-CTGGGTTCCCTGTTCTGTGT	312
Myogenin	5′-GGGGACCCCTGAGCATTGTCC	5′-TGGACATCAGGACAGCCCCAC	613
α3-Integrin	5′-CAAGGACGACTGTGAACGGA	5′-CCTGCACCGTGTACCCAATA	470
β1-Integrin	5′-TCAGACTTCCGCATTGGCTT	5′-TGTGCCCACTGCTGACTTAG	546
Cd9	5′-GTACCATGCCGGTCAAAGGA	5′-CACAGCAGTCCAACGCCATA	459
M-cadherin	5′-GCCCTGCTTCACCCTTTTTG	5′-CGTGCATGGCTCACGTTAAT	316
N-cadherin	5′-AGCCCGGTTTCACTTGAGAG	5′-TCCGTGACAGTTAGGTTGGC	531
Myomaker
(Tmem8c)	5′-ATCGCTACCAAGAGGCGTT	5′-CACAGCACAGACAAACCAGG	117
Arf6	5′-CGGGAACAAGGAAATGCGGA	5′-GGTCCCGGTGTAGTAATGCC	217
Brag2	5′-GACCCCAACAAACCCCAGAA	5′-AGGGCTGAAATTCCACATTGC	568
Dock5	5′-TGGAGATGTATGAGGGCTGGT	5′-CCGCATCCTCCCCAATGTTA	606
Fak1 (Ptk2)	5′-GTGTCAAGCTTCAGCCCCAGGA	5′-TGGCCGTGTCTGCCCTAGCA	452
Trio	5′-CTTCTTCCGATCCGGGTTTC	5′-GGCAACTCCTTTTTGGCGAG	597
Wasp (Was1)	5′-GCAGTGGTGCAGTTGTATGC	5′-ATTGGGACCATTTGGAGCATCT	339
Rock2M	5′-CGACTCATATCGCCCGAGG	5′-AGTCGTACCTCCCTGTCTGT	258
S17	5′-ATCCCCAGCAAGAAGCTTCGGAACA	5′-TATGGCATAACAGATTAAACAACTC	302

Protein immunoblotting.

Whole cell protein lysates were prepared as described elsewhere (12). Aliquots (15 μg/lane) were resolved by SDS-PAGE (10% separating gel), transferred to Immobilon-FL membranes, and blocked with a 50% solution of AquaBlock. Membranes were incubated sequentially with primary antibodies for 16 h at 1:1,000 dilution, except for α-tubulin at 1:10,000 dilution, and with secondary antibodies for 1.5 h at 1:5,000 dilution (35). Images were captured using the Odyssey imaging system and version 3.0 analysis software (LI-COR, Lincoln, NE). All immunoblotting was performed at least three times using protein lysates derived from independent experiments, and representative results are depicted in Fig. 3.

Fig. 3.

Muscle and signaling protein expression after removal of the p38 inhibitor SB202190. Protein expression by immunoblotting for MyoD, myogenin, Mef2C, troponin-T, myosin heavy chain (MHC), phosphorylated (p)-p38 and total p38, p-Erk and total Erk, p-Akt (pAkt^S308) and total Akt, p-mammalian target of rapamycin (mTor) and total mTor, and α-tubulin by immunoblotting during incubation of C2 myoblasts in DM plus SB202190 and R3-IGF-I or after removal of SB202190. Molecular mass markers are indicated at right. Images are representative of ≥3 independent experiments.

Preparation of tryptic peptides for mass spectrometry.

C2 myoblasts at ∼95% of confluent density were incubated in DM with SB202190 (5 μM) and R3-IGF-I (1 nM). After 48 h, cells were washed with DMEM, and medium was replaced with DM plus SB202190 and R3-IGF-I or with DM and R3-IGF-I. Protein lysates were collected 0, 4, 12, and 24 h later. Cells were washed twice with PBS, ammonium bicarbonate (50 mM) was added, and cells were scraped from the plate and lysed by five passages through a 22-gauge needle. Each sample was then sonicated using a probe (and model 60 sonic dismembrator, Thermo Fisher Scientific) for a total of three rounds of 5-s bursts at a 4-W setting, each interspersed with 15-s incubations on ice. Protein concentrations were determined using a BCA assay with bovine serum albumin as standard. Lysates were dried in 100-μg aliquots and resuspended in 50 μl of 8 M deionized urea, 1 M Tris (pH 8.5), 8 mM CaCl₂, and 0.2 M methylamine. DTT was added to 15 mM, and samples were incubated at 50°C for 30 min. After they were cooled to 15°C, the samples were alkylated by addition of iodoacetamide (40 mM) for 15 min in darkness and quenched with additional DTT (30 mM) for 15 min in darkness. Samples were diluted to 2 M urea, and 4 μg of sequencing-grade trypsin were added. After overnight digestion at 37°C, samples were acidified by addition of 1.5% trifluoroacetic acid, applied to Sep-Pak Light C18 cartridges, and dried by vacuum centrifugation in 25-μg aliquots. Each digest was dissolved in 25 μl of 100 mM triethylammonium bicarbonate buffer and labeled using 200 μg of freshly prepared TMT10plex reagent dissolved in anhydrous ethanol according to the manufacturer-suggested protocol, except peptide and reagent amounts were decreased by a factor of 4. After they were quenched using hydroxylamine, labeled samples were pooled and dried by vacuum centrifugation in preparation for mass spectrometric analysis.

Mass spectroscopy.

TMT-labeled peptides were separated by two dimensions of online reverse-phase chromatography using a Dionex NCS-3500RS UltiMate RSLCnano ultraperformance liquid chromatography (UPLC) system for sample loading and second-dimension reverse-phase separation and a Dionex NCP-3200RS UltiMate RSLCnano UPLC system for dilution of the first-dimension eluent. Pooled TMT-labeled digest (20 μg) was injected for 10 min onto a NanoEase 5-μm XBridge BEH130 C18 300 μm × 50 mm column (Waters) at 3 μl/min in a mobile phase containing 10 mM ammonium formate (pH 10) and 2% acetonitrile. Peptides were eluted by sequential injection of 20-μl volumes of 8, 14, 17, 20, 23, 26, 30, 50, and 90% acetonitrile in 10 mM ammonium formate (pH 10) at 3 μl/min flow rate. Eluted peptides were then diluted at a T-connector with a mobile phase containing 0.1% formic acid at a 12 μl/min flow rate. Peptides were delivered to an Acclaim PepMap 100 μm × 2 cm NanoViper C18, 5-μm trap on a switching valve. After 10 min of loading, the trap column was switched online to a PepMap RSLC C18, 2-μm, 75 μm × 25 cm EasySpray column (Thermo Fisher Scientific). Peptides were then separated at low pH in the second dimension using a 7.5–30% acetonitrile gradient over 92 min in mobile phase containing 0.1% formic acid at a 300 nl/min flow rate. Tandem mass spectrometry data were collected using an orbitrap Fusion Tribrid instrument configured with an EasySpray NanoSource (Thermo Fisher Scientific). Survey scans were performed in the orbitrap mass analyzer, and data-dependent MS2 scans were performed in the linear ion trap using collision-induced dissociation following isolation with the instrument's quadrupole. Reporter ion detection was performed in the orbitrap mass analyzer using MS3 scans following synchronous precursor isolation in the linear ion trap and higher-energy collisional dissociation in the ion-routing multipole.

TMT data analysis.

RAW instrument files were processed with Proteome Discoverer version 1.4.1.14 (Thermo Fisher Scientific) using SEQUEST HT software and a Swiss Prot canonical mouse protein database (version 2014.05, Swiss Bioinformatics Institute, Geneva, Switzerland). Searches were configured with static mass modifications for the TMT reagents (+229.163 for peptide NH₂ terminus and lysines) and alkylated cysteines (+57.021), variable oxidation of methionine (+15.995), parent ion tolerance of ±1.25 Da, fragment ion tolerance of 1.0 Da, monoisotopic masses, and trypsin cleavage (≤2 missed cleavages). Searches used a reverse-sequence decoy strategy to control peptide false discovery (11). All identifications were validated by Percolator software (20). Only peptides with q values ≤0.05 were accepted. TMT reporter ion quantification was performed using an integration tolerance of 20 ppm of the most confident centroid in MS3 scans. Search results and TMT reporter ion intensities were exported as text files and processed with in-house scripts. A median reporter ion intensity cutoff of 2,000 was used to reject low-quality peptides, and all reporter ion intensities for unique peptides matched to each respective protein were summed to create total protein intensities. A minimum of two peptides contributing to the protein total were also used to improve data quality.

For subsequent analysis, summed protein intensity values of <10,000 obtained from lysates of cells incubated in DM with SB202190 and IGF-I for 48 h were eliminated from further consideration. Results for the remaining 2,690 proteins were normalized by determination of the ratio of protein intensity at different time points and treatments to protein intensity from lysates of cells incubated for 48 h in DM with SB202190 and IGF-I. The 100 proteins with the highest and lowest ratios in each comparison were studied further (see Fig. 5C and Tables 3 and 4). For Fig. 5, A and B, the following ratios were analyzed: 1) cells incubated in DM and IGF-I for 24 h after removal of SB202190 vs. cells exposed to SB202190 plus IGF-I for 48 h, 2) cells incubated in DM and IGF-I for 24 h after removal of SB202190 vs. cells in DM plus SB202190 and IGF-I for all 72 h, and 3) cells in DM and IGF-I for 48 h vs. cells in DM plus SB202190 and IGF-I for 48 h (for a complete list of identified proteins, raw summed reporter ion intensities, and normalized data used to calculate protein abundance ratios, see Supplemental Table S1 in Supplemental Material for this article available online at the Journal website). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (54) via the PRIDE partner repository with the dataset identifier PXD002413.

Fig. 5.

Changes in protein expression during muscle cell fusion. A and B: Venn diagrams showing relative alterations in protein abundance as assessed by mass spectroscopy of protein lysates from C2 myoblasts incubated under various conditions. Specific pair-wise comparisons depicted by each circle are listed, as are the number of unique and overlapping proteins identified from the top 100 molecules in each analysis. A: proteins that increased in abundance after removal of SB202190 or in its absence. B: proteins that decreased in abundance. C: heat map showing relative log₁₀ scale changes in protein abundance under different experimental conditions (+, continued presence of SB in cell culture medium; −, removal of SB202190 at the 48-h time point). Displayed are the 56 proteins present in all 3 pairwise comparisons that increased in abundance and the 19 proteins that decreased. Abundance values were normalized based on the protein amounts measured in cells incubated with SB202190 for 48 h. Scale bar illustrates dynamic range of relative protein concentrations.

Table 3.

Proteins upregulated after removal of SB202190

Protein	Description
Acta1	Actin, α1, skeletal muscle
Acta2	Actin, α2, smooth muscle, aorta
Actn2	Actinin, α2
Ak1	Adenylate kinase 1
Akr1b8	Aldo-keto reductase family 1, member B8
Atp2a1	ATPase, Ca2+-transporting, cardiac muscle, fast-twitch 1
Bicd2	Bicaudal D homolog 2
Cap2	CAP, adenylate cyclase-associated protein, 2
Ccdc134	Coiled-coil domain containing 134
Ckb	Creatine kinase, brain
Ckm	Creatine kinase, muscle
Cox6b1	Cytochrome c oxidase, subunit VIb polypeptide 1
Ctage5	CTAGE family, member 5
Dclk1	Double cortin-like kinase 1
Ehbp1l1	EH domain-binding protein 1-like 1
Eno3	Enolase 3, β, muscle
Igf2	Insulin-like growth factor II
Jph2	Junctophilin 2
Klhl41	Kelch-like protein 41
Kif1b	Kinesin family member 1B
Ldb3	LIM domain-binding 3
Macf1	Microtubule-actin cross-linking factor 1
Murc	Muscle-related coiled-coil protein
Mybph	Myosin-binding protein H
Myh1	Myosin, heavy polypeptide 1
Myh3	Myosin, heavy polypeptide 3
Myl1	Myosin, light polypeptide 1
Myl4	Myosin, light polypeptide 4
Myl6b	Myosin, light polypeptide 6B
Mylpf	Myosin light chain, phosphorylatable, fast skeletal muscle
Myom1	Myomesin 1
Myom3	Myomesin family, member 3
Naca	Nascent polypeptide-associated complex-α polypeptide
Nexn	Nexilin
Pacsin3	PKC and casein kinase substrate in neurons 3
Pdlim4	PDZ and LIM domain 4
Pfkm	Phosphofructokinase, muscle
Pgam2	Phosphoglycerate mutase 2
Polr2 h	Polymerase (RNA) II (DNA-directed) polypeptide H
Pygm	Muscle glycogen phosphorylase
Smyd1	SET and MYND domain-containing 1
Sntb1	Syntrophin, basic 1
Srl	Sarcalumenin
Stim1	Stromal interaction molecule 1
Synpo2	Synaptopodin 2
Synpo2l	Synaptopodin 2-like
Tnnc1	Troponin C, cardiac/slow skeletal
Tnnc2	Troponin C2, fast
Tnni1	Troponin I, skeletal, slow 1
Tnni2	Troponin I, skeletal, fast 2
Tnnt2	Troponin T2, cardiac
Tnnt3	Troponin T3, skeletal, fast
Tpm1	Tropomyosin 1, α
Tpm2	Tropomyosin 2, β
Trim72	Tripartite motif-containing 72
Unc45b	Unc-45 homolog B

Table 4.

Proteins downregulated after removal of SB202190

Protein	Description
A2m	α2-Macroglobulin
Aebp1	AE-binding protein 1
Atp9b	ATPase, class II, type 9B
Casp6	Caspase 6
Flnb	Filamin, β
Ftl1	Ferritin light chain 1
Impact	Imprinted and ancient
Lars2	Leucyl-tRNA synthetase, mitochondrial
Marcks	Myristoylated alanine-rich PKC substrate
Marcksl1	MARCKS-like 1
Mfge8	Milk fat globule-EGF factor 8 protein
Ndrg1	N-myc downstream-regulated gene 1
Ovca2	Candidate tumor suppressor in ovarian cancer 2
Rnf114	Ring finger protein 114
Scpep1	Serine carboxypeptidase 1
Sdcbp	Similar to syntenin; syndecan-binding protein
Slc16a1	Solute carrier family 16 (monocarboxylic acid transporters), member 1
Tpm4	Tropomyosin 4
Tsc22d1	TSC22 domain family, member 1

Statistical analysis.

Values are means ± SD, unless otherwise indicated. Statistical analysis was performed using paired Student's t-test. Results were considered statistically significant when P ≤ 0.01.

RESULTS

IGF-I promotes muscle differentiation in the presence of a p38 MAP kinase inhibitor.

Previous studies have shown that inhibition of p38 MAP kinase activity can diminish the rate and extent of skeletal muscle differentiation in vitro (9, 32, 55, 57). We find similar effects with the selective p38α/β chemical inhibitor SB202190, which, when added to DM, reduced the number of elongated troponin-T-expressing myocytes that developed in C2 myoblasts cultured over a 72-h incubation period (Fig. 1). Addition of the IGF-I analog R3-IGF-I with SB202190 increased differentiation by ∼3.5-fold when assessed by the area occupied by troponin-T-containing cells (Fig. 1C). Yet, although these cells expressed muscle proteins and were both elongated and aligned, they were almost entirely mononucleated (Fig. 1D). Thus, IGF-I-mediated signaling is able to restore muscle differentiation when p38 MAP kinase activity is blocked but is not able to promote myocyte fusion.

Fig. 1.

Insulin-like growth factor (IGF)-I restores muscle differentiation blocked by the p38 inhibitor SB202190. A: experimental plan. Confluent C2 myoblasts were incubated in differentiation medium (DM) for up to 72 h in the absence or presence of SB202190 (5 μM) ± R3-IGF-I (1 nM). B: immunocytochemistry after 24, 48, and 72 h in DM for troponin-T (red) and Hoechst nuclear stain (blue). SB, SB202190. Magnification ×100 (scale bars = 100 μm). C: area of troponin-T after 72 h in DM. Values are means ± SD. *P < 0.00001, **P < 0.0000003, ***P < 0.0000002. D: immunocytochemistry for troponin-T (red) and myogenin (green) and staining for nuclei with Hoechst 33258 (blue). Magnification ×200 (scale bar = 50 μm).

Myofiber formation does not occur in the presence of a p38 MAP kinase inhibitor.

To directly address the possibility that p38 signaling activity is needed for muscle cell fusion, C2 myoblasts were incubated in DM with SB202190 and R3-IGF-I for 44 h, at which time the inhibitor was either maintained in the medium or removed (Fig. 2A). In the continued presence of SB202190 and IGF-I, extensive differentiation was noted, but very few multinucleated myofibers were detected by immunocytochemistry by 74 h (Fig. 2, B and C). In contrast, elimination of SB202190 led to robust formation of multinucleated myotubes over the subsequent 30 h (Fig. 2, B and C). As recorded by live-cell imaging, progressive fusion was observed beginning several hours after the inhibitor was removed, and ∼90% myotube formation was attained ∼30 h later (Fig. 2D; see Supplemental Movies S1 and S2 in Supplemental Material for this article) vs. ∼20% in the presence of SB202190 (Fig. 2D). Thus, SB202190 prevents muscle fusion, even in the presence of the prodifferentiation activity of IGF-I, and its subsequent removal enables fusion. As a consequence, treatment with SB202190 plus IGF-I allows the separation of muscle differentiation from fusion.

Fig. 2.

IGF-I promotes muscle differentiation, but not myocyte fusion, in the presence of the p38 inhibitor SB202190. A: experimental plan. Confluent C2 myoblasts were incubated in DM with SB202190 (5 μM) ± R3-IGF-I (1 nM). After 44 h, medium was replaced, and new DM was added ± SB202190 for an additional 30 h. B: immunocytochemistry for troponin-T (red) and myogenin (green). Magnification ×100 (scale bars = 100 μm). C: immunocytochemistry for troponin-T (red) and myogenin (green) and staining for nuclei with Hoechst 33258 (blue). Magnification ×200 (scale bars = 50 μm). D: changes in the extent of myocyte fusion over time during incubation ± SB during live-cell imaging.

Signaling protein expression during fusion.

We next examined protein expression in C2 muscle cells after SB202190 was maintained in the medium or removed from the medium, beginning at 44 h after onset of incubation in DM. There was little alteration in levels of muscle transcription factors, MyoD, myogenin, and Mef2C during the subsequent 30 h, although an increase was noted in muscle structural proteins, troponin-T and myosin heavy chain, beginning at 8–12 h after removal of SB202190 (Fig. 3). In contrast, a sustained decline was seen in the amount of p-p38^T180/Y182 when SB202190 was washed out (Fig. 3). This is consistent with loss of feedback inhibition when SB202190 is removed. Only transient alterations were observed in other signaling molecules, including p-Akt^T308, p-mTor^S2448, and p-Erk^T202/Y204 (Fig. 3), and steady-state levels of their total proteins were constant.

Limited alteration in the expression of mRNAs and proteins involved in muscle fusion.

We next assessed a cohort of molecules that have been implicated in muscle cell fusion (1, 23, 42) by examining changes in their gene and protein expression at different times after removal of SB202190 from DM. Results of mRNA abundance using RT-PCR and protein levels using quantitative mass spectroscopy of the same fusion-associated molecules are shown in Fig. 4A and Table 2. Steady-state expression of transcripts for transmembrane proteins, β₁-integrin, Cd9, and N-cadherin, were unchanged, as were levels of fusogenic intracellular transducers, Arf6, Trio, Was1, and the muscle isoform of Rho-associated protein kinase 2 Rock2M (Fig. 4A) (42). We also measured several other mRNAs encoding fusion-associated proteins that were not found consistently in our mass spectrometry data (see below), including α₃-integrin, M-cadherin, myomaker (Tmemb8C) (34), Brag2, Dock5, and Fak1 (1). Except for α₃-integrin, which was transiently increased, we detected no changes in abundance of these latter mRNAs when comparing cells incubated with and without SB202190 for up to 30 h (Fig. 4B).

Fig. 4.

Limited alterations in expression of genes encoding proteins involved in muscle cell fusion. Time course of accumulation of various mRNAs as measured by RT-PCR during incubation of C2 myoblasts in DM plus SB202190 (5 μM) and R3-IGF-I (1 nM) for up to 74 h or after their removal during the last 30 h. A: transcripts assessed include those encoding the muscle transcription factors MyoD and myogenin, transmembrane proteins implicated in myocyte fusion (β₁-integrin, Cd9, and N-cadherin), and cytoplasmic and cytoskeletal proteins involved in fusion (Arf6, Trio, Was1, and Rock2M). B: transcripts measured include transmembrane proteins implicated in fusion (α₃-integrin, M-cadherin, and myomaker), cytoplasmic and cytoskeletal proteins involved in fusion (Brag2, Dock5, and Fak1), and control mRNA S17.

Table 2.

Relative expression of fusion-associated proteins

		Ratio of SB Removed to SB Maintained
Protein	Category	4 h	12 h	24 h
N-cadherin	Transmembrane	0.97	0.97	1.00
Cd9	Transmembrane	1.06	0.89	1.12
β1-Integrin	Transmembrane	0.92	1.03	1.01
Arf6	Small GTPase	0.98	0.93	1.13
Trio	GTPase exchange factor	1.10	1.05	1.18
Was1	Actin-interacting	0.99	0.82	1.03
Rock2	Protein kinase	1.00	0.99	0.95
N-cam	Transmembrane	0.95	1.08	1.31
Kindlin2	Transmembrane	0.98	1.02	1.33
Cdc42	Small GTPase	0.95	1.13	0.98
Rac1	Small GTPase	1.03	1.09	0.99
Dock1	GTPase exchange factor	0.90	0.89	1.06
Nap1	Actin-interacting	0.91	0.95	1.00

SB, SB202190.

To more broadly analyze changes in protein abundance, we performed quantitative mass spectroscopy on paired protein lysates at three intervals following 48 h of incubation in DM with SB202190 and IGF-I: 4, 12, and 24 after maintenance or removal of SB202190. We also collected protein lysates from cells incubated in DM plus IGF-I for 48 h. From these lysates, we identified 2,690 proteins whose levels could be tracked at different time points. Analysis of the expression of fusion-associated proteins whose mRNA levels were unchanged, including β₁-integrin, Cd9, N-cadherin, Arf6, Trio, Was1, and Rock2, as well as several other molecules implicated in muscle fusion [N-cam, kindlin2, Cdc42, Rac1, Dock1, and Nap1 (1)], yielded at most minimal alterations in abundance following washout compared with cells that were maintained in SB202190 (Table 2). Although not all proteins previously characterized to be involved in muscle fusion were present in the mass spectroscopy data, these results show that SB202190 does not prevent or alter the expression of a substantial number of fusion-associated molecules.

Global characterization of protein expression during muscle cell fusion.

We next performed a series of unbiased analyses of the mass spectroscopy data in search of proteins that changed in abundance in the presence or absence of SB202190. We used three partially overlapping screening approaches and focused on the 100 proteins in each screen that showed the largest relative increase or decrease compared with an SB202190 control sample. For these analyses, pair-wise comparisons were made between 1) cells incubated in DM and IGF-I for 24 h after removal of SB202190 vs. those exposed to SB202190 plus IGF-I for 48 h, 2) cells incubated in DM and IGF-I for 24 h after removal of SB202190 vs. cells incubated in DM plus SB202190 and IGF-I for all 72 h, and 3) cells incubated in DM and IGF-I for 48 h vs. cells incubated in DM plus SB202190 and IGF-I for 48 h.

From these analyses, we identified 82 proteins that increased in abundance after removal of SB202190 and were present in at least two of the three pair-wise comparisons (Fig. 5A, Table 3). This included 56 proteins identified in all three screens (Fig. 5A), also plotted in the heat map in Fig. 5C, which illustrates their relative change in abundance under different experimental conditions with and without SB202190. Nearly half of these molecules are components of sarcomeres or bind to sarcomeric proteins (Table 3). These molecules are important for muscle contraction, a major function of multinucleated myofibers, but their possible roles in myocyte fusion are unknown.

We identified 48 proteins that decreased in abundance in cells incubated without SB202190 in at least two comparisons, including 19 proteins present in all three screens (Fig. 5, B and C, Table 4). Noteworthy in this group are multiple cytoskeletal proteins (e.g., Flnb, Marcks, Marcksl1, and Sdcbp) (Table 4), suggesting a cellular function potentially sustained in differentiating muscle cells in the presence of SB202190.

DISCUSSION

The role of p38 MAP kinases in skeletal muscle has been recognized for over 15 years, since it was shown that specific chemical inhibitors of these enzymes reversibly blocked differentiation in culture (9, 57). Other studies found that p38 kinase activity increased during muscle differentiation (55, 57) and that forced activation of this pathway could stimulate differentiation (55) and promote myotube formation in otherwise nonfusing rhabdomyosarcoma cells in culture (46). Signaling by p38α was found to enhance the myogenic transcriptional program through direct and indirect interactions with MyoD and Mef2 (21, 30, 55–57, 59) and through facilitation of chromatin opening at muscle gene promoters (48, 49). Despite these observations, the mechanisms that regulate p38 MAP kinase activity and the specific processes that require p38 in differentiating myoblasts remain unknown (31).

Here we show that sustained IGF-I-stimulated signaling can overcome impairments caused by the p38-specific inhibitor SB202190 and drive muscle differentiation that is otherwise blocked. Despite substantial expression of muscle genes and proteins under these experimental conditions and elongation and alignment of the differentiating myocytes, muscle cells incubated with SB202190 and IGF-I did not fuse but, rather, remained primarily mononucleated (Figs. 1 and 2). Robust and extensive fusion could be triggered by removal of SB202190 from the medium, such that, by ∼30 h later, ∼90% of the cells were in multinucleated myofibers (Fig. 2, C and D). Thus our data indicate that p38 MAP kinase activity is critical for muscle cell fusion.

Since it was not obvious how inhibition of p38 prevented formation of multinucleated myotubes, we performed a series of experiments to identify molecules that changed in abundance when fusion was initiated by removal of SB202190 from the medium. Analysis of known fusion-linked proteins demonstrated that many were expressed in myoblasts incubated in DM with SB202190 and IGF-I but that their abundance at the mRNA or protein level did not change appreciably in fusing myofibers (Fig. 4, Table 2).

We next sought to use a nondirected approach to identify proteins whose quantities either increased or decreased after removal of SB202190 from DM, assuming that at least some of these molecules would be linked in a positive or negative way to myocyte fusion. The vast majority of the 56 proteins found to be upregulated in multiple paired quantitative proteomics experiments were either enriched or predominantly expressed in striated muscle, thus validating our screen. At least half of these molecules are components of myofibrils or are known to interact with one or more of these myofibrillar proteins (Fig. 5, Table 3). These proteins are clearly critical for the contractile functions of multinucleated myotubes, but their involvement in fusion is unclear.

Recent studies in Drosophila muscle have established roles for actin cytoskeletal rearrangement in facilitating the cell membrane proximity required for fusion and have shown that nonmuscle myosin II is a key “motor” protein responsible for generating sufficient contractile mechanical stresses in the cytoskeleton through the force of actin polymerization (8, 22). Myh9, Myh10, and Myg14 are the mammalian orthologs for the heavy chain of Drosophila myosin II (4, 53). All three of these nonmuscle myosins were present in our cells, but their abundance did not change after removal of SB202190 from DM (data not shown).

Expression of two other muscle-enriched proteins, Smyd1, a member of a family of SET-MYND domain-containing lysine methyltransferases (10), which are critical for cardiac and skeletal muscle development (18), and Naca, whose muscle-specific isoform, Sknac, is the major binding partner of Smyd1 in striated muscle (41), increased after removal of SB202190 from DM. Sknac knockout mice display impaired muscle growth and regeneration (41). In addition, loss of Sknac or Smyd1 in zebrafish leads to disordered myofibrils (26, 27). Both molecules might be good targets for future studies of the mechanics of muscle fusion and myofiber formation.

In contrast to results obtained among proteins enriched in cells after removal of SB202190, most of the 19 molecules with decreased abundance were not predominantly expressed in striated muscle but tended to be relatively ubiquitous (Fig. 5, Table 4). Unlike the upregulated proteins, these molecules did not cluster into any specific categories; however, several are found in the cytoskeleton and, thus, reveal little about possible mechanisms that prevent muscle fusion in cells in which p38 MAP kinase activity is inhibited.

In summary, through use of IGF-I plus the p38α/β MAP kinase inhibitor SB202190, we have developed a mammalian muscle cell model in which we have separated myoblast differentiation from myocyte fusion. Analysis of this model has yielded a relatively large number of contractile proteins and related molecules apparently upregulated by p38 activity during formation of multinucleated myofibers. The potential roles of any of these proteins in muscle cell fusion remain to be defined, and this model may prove useful in determining how the regulatory steps mediated by signaling pathways become transduced into cellular outcomes (19).

GRANTS

Mass spectrometric analysis was performed by the Oregon Health and Science University Proteomics Shared Resource with partial support from National Institutes of Health Core Grants P30 EY-010572, P30 CA-069533, and S10 OD-012246. These studies were supported by National Institutes of Health Grants R01 DK-042748-25 (to P. Rotwein) and T32 CA-106195 (supporting S. M. Gross).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

S.G., S.M.G., L.L.D., and P.R. developed the concept and designed the research; S.G., S.M.G., and J.E.K. performed the experiments; S.G., S.M.G., L.L.D., J.E.K., and P.R. analyzed the data; S.G., S.M.G., L.L.D., J.E.K., and P.R. interpreted the results of the experiments; S.G., S.M.G., and P.R. prepared the figures; S.G. and P.R. drafted the manuscript; S.M.G. and P.R. edited and revised the manuscript; P.R. approved the final version of the manuscript.