Autocrine Growth Factor Signaling by Insulin-like Growth Factor-II Mediates MyoD-stimulated Myocyte Maturation*

Skeletal muscle differentiation, maturation, and regeneration are regulated by interactions between intrinsic genetic programs controlled by myogenic transcription factors, including members of the MyoD and MEF2 families, and environmental cues mediated by hormones and growth factors. Insulin-like growth factors (IGFs) also play key roles in muscle development, and in the maintenance and repair of mature muscle, but their mechanisms of interaction with other muscle regulatory networks remain undefined. To evaluate the potential interplay between MyoD and IGF signaling pathways, we have studied muscle differentiation in C3H 10T1/2 fibroblasts acutely converted to myoblasts by quantitative infection with a recombinant adenovirus encoding mouse MyoD. In these cells, IGF-II gene and protein expression are induced as early events in differentiation, and the IGF-I receptor and downstream signaling molecules, including Akt, are rapidly activated. Interference with IGF-II production by a tetracycline-inhibited adenovirus expressing an IGF-II cDNA in the antisense orientation reversibly inhibited both production of muscle-specific structural proteins and myocyte fusion to form multinucleated myotubes. Similar results were achieved with a tetracycline-inhibited adenovirus expressing dominant-negative Akt. Our observations identify a robust autocrine amplification network in which MyoD enhances the later steps in muscle differentiation by induction of a locally acting growth factor.

Skeletal muscle differentiation, maturation, maintenance, and repair require ongoing cooperation and coordination between an intrinsic regulatory program controlled by myogenic transcription factors and signaling pathways activated by hormones and growth factors (1,2). The basic helix-loop-helix (bHLH) 1 transcription factors of the MyoD family (MyoD, Myf-5, myogenin, and MRF4) are central regulators of myogenesis in vivo and in cell culture (3,4). These proteins bind as heterodimers with other widely expressed bHLH transcription factors to DNA control regions termed E boxes that are often found in the promoters of muscle-specific genes (4). Subsequent interactions with transcriptional co-activators are necessary for gene activation (5,6). As established by gene knock-out studies in mice, MyoD and Myf-5 act redundantly at an early step in myoblast specification. Mice expressing either factor alone form muscle but do not if both proteins are absent (7). MyoD also is required for normal muscle regeneration in the adult (8). Myogenin acts downstream of MyoD/Myf-5 to promote differentiation (9,10), while MRF4 plays a more limited role in muscle formation in vivo (reviewed in Ref. 11).
Insulin-like growth factors (IGF-I and -II) also play key roles in normal muscle development in the embryo (12,13) and are important for coordinating muscle regeneration and re-innervation following injury (14,15). IGF action additionally may be critical for sustaining muscle mass during aging (16 -19). In muscle and in other cell types the actions of both IGFs are mediated by the IGF-I receptor (IGF-IR), a ligand-stimulated membrane-spanning tyrosine protein kinase that activates several intracellular signaling pathways through a series of adaptor molecules (20). IGF action also is modified through high affinity interactions with a series of extracellular IGF-binding proteins, IGFBP-1 through -6 that limit access to the IGF-IR, but also enhance growth factor half-life (21).
In contrast to differences observed among bHLH proteins in vivo, tissue culture studies have demonstrated similar functions for these transcription factors. In culture, each protein has been found to direct a range of cell types toward the myoblast lineage, to promote cell cycle arrest, and to stimulate differentiation (1,3), in part through cooperation with members of the MEF2 family of transcriptional factors, which act as accessory regulators of muscle gene expression and differentiation (2). Here using C3H 10T1/2 fetal fibroblasts as a model, we identify another mechanism by which MyoD enhances its myogenic actions, through induction of IGF-II gene and protein expression early in differentiation. Secreted IGF-II in turn activates the IGF-IR and downstream signaling molecules, including Akt. As inhibition of IGF-II expression or impairment of Akt prevents production of muscle structural proteins and blocks formation of multinucleated myofibers, these results define an autocrine amplification pathway by which myogenic bHLH proteins stimulate the later events of muscle differentiation through production of a locally acting growth factor.
(Madison, WI). The BCA protein assay kit was from Pierce, and nitrocellulose was from Osmonics (Westborough, MA). Restriction enzymes, buffers, ligases, and polymerases were purchased from Roche Applied Sciences, BD Biosciences (Clontech), and Fermentas (Hanover, MD). Reagents for enhanced chemifluorescence were from Amersham Biosciences. Several monoclonal antibodies were purchased from the Developmental Studies Hybridoma Bank (Iowa City, IA), including F5D (anti-myogenin, W. E. Wright), MF20 (anti-myosin heavy chain (MHC), D. A. Fischman), and CT3 (anti-troponin T, J. J.-C. Lin). A monoclonal antibody to MyoD was from BD Biosciences (Pharmingen, San Diego, CA). Polyclonal antibodies to Akt and phospho-Akt (Ser 473 ) were from Cell Signaling Technology (Beverly, MA), and the polyclonal antibody to IGF-II was purchased from Abcam, Ltd. (Cambridge, UK). A monclonal antibody to phosphotyrosine was from Santa Cruz Biotechnology (Santa Cruz, CA), as was a polyclonal antibody to the ␤ subunit of the IGF-I receptor. Antibody conjugates were purchased from Molecular Probes (Eugene, OR): goat anti-mouse IgG 1 -Alexa 488, goat anti-mouse IgG 2b -Alexa 594, anti-rabbit IgG-alkaline phosphatase, and anti-mouse IgGalkaline phosphatase. The AdEasy adenoviral recombinant kit was from Q-BIO Gene (Carlsbad, CA). All other chemicals were reagent grade and were purchased from commercial suppliers.
Cell Culture-C3H 10T1/2 mouse embryonic fibroblasts (ATCC catalog number CCL226) were incubated on gelatin-coated tissue culture dishes in growth media (DMEM with 10% heat-inactivated fetal bovine serum and 10% newborn calf serum) at 37°C in humidified air with 5% CO 2 , until they reached 50% of confluent density for acute infection with recombinant adenoviruses. Differentiation was initiated 1 day later after cells reached ϳ95% of confluent density after washing with phosphate-buffered saline by addition of differentiation medium (DM), consisting of DMEM plus 2% horse serum (22). In some experiments DM consisted of DMEM plus 0.5% horse serum.
Construction and Use of Recombinant Adenovirus-A FLAG epitope tag followed by a stop codon and XbaI site were added to the 3Ј end of the coding region of mouse MyoD by polymerase chain reaction. The modified cDNA was sequenced, digested with SalI and XbaI restriction endonucleases, and ligated into the pShuttle:CMV vector. A dominantnegative Akt (Akt DN ) was prepared by modifying a human Akt-1 cDNA with a T7 epitope tag at its 3Ј end (a gift from Dr. Richard Roth, Stanford University School of Medicine) by mutating codons for lysine 179, threonine 308, and serine 473 to alanines using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Each mutation was verified by restriction enzyme mapping and by DNA sequencing. The Akt DN cDNA was then subcloned via SalI and XbaI restriction endonuclease sites into a modified pShuttle plasmid containing a tetracycline-regulated promoter (22), as was the coding region of mouse IGF-II in the antisense orientation (23). All recombinant adenoviruses were generated and isolated via a protocol supplied by Q-BIO Gene. The adenovirus encoding a tetracyline-inhibited transactivator (Ad-tTA) has been described (22). A recombinant adenovirus encoding ␤-galactosidase (Ad-␤-Gal) was a gift from Dr. J. Molkentin, University of Cincinnati School of Medicine. All viruses were purified on discontinuous cesium chloride gradients and titered by optical density.
For infections, recombinant adenoviruses (Ad-tTA at a multiplicity of infection of 125, others at 250) were diluted in DMEM plus 2% fetal bovine serum, filtered through a Gelman syringe filter (0.45 M), and added to cells at 37°C for 120 min. After addition of an equal volume of DMEM with 20% each of fetal bovine serum and newborn calf serum, cells were incubated for a further 24 h and then were placed in DM. Under these conditions, ϳ90% of cells were infected.
Immunocytochemistry-Cells were fixed in 4% paraformaldehyde for 15 min at 20°C and permeabilized with a 50:50 mixture of methanol and acetone for 2 min before blocking in 0.25% normal goat serum for Ͼ1 h at 20°C. Primary antibodies diluted in blocking buffer were added for 16 h at 4°C (anti-MHC, 1:250 dilution, anti-myogenin, 1:250 dilution). After a washing step, cells were incubated for 2 h at 20°C in goat anti-mouse IgG 2b -Alexa 594 (red) and goat anti-mouse IgG 1 -Alexa 488 (green), each diluted to 1:1000 in blocking buffer. Images were captured with a Roper Scientific Cool Snap FX CCD camera attached to a Nikon Eclipse T300 fluorescent microscope using IP Labs 3.5 software.
Here we show that Akt enzymatic activity also is induced as an early event in differentiation in fibroblasts converted to myoblasts after acute infection with a recombinant adenovirus encoding mouse MyoD (Ad-MyoD) and that active Akt is needed for the later events in differentiation that culminate in myotube formation. Illustrated in Fig. 1 are results of time course experiments using 10T1/2 fibroblasts infected 1 day earlier with Ad-MyoD at an m.o.i. in which ϳ90% of cells express MyoD (data not shown) and subsequently incubated in DM. As seen in Fig. 1A, Ad-MyoD-infected 10T1/2 cells underwent rapid and extensive muscle differentiation, with progressive expression of myogenin and MHC, and formation of large multinucleated myotubes within 2 days. Similarly robust muscle-specific protein expression was observed by immunoblotting (Fig. 1B). Myogenin and MHC were induced in Ad-MyoD-infected cells but were not detected in fibroblasts infected with an adenovirus encoding ␤-galactosidase (Ad-␤-Gal). Also seen in Fig. 1B is evidence of activation of Akt beginning by 16 h after addition of DM, as indicated by a progressive increase in phosphorylation on serine 473. A similarly large induction of Akt enzymatic activity was observed, as shown in Fig. 1C. In Ad-MyoD infected cells, Akt kinase activity was stimulated by ϳ20-fold by 1 day in DM, was increased by over 30-fold by 2 days, and was maintained at high levels for at least 3 days. Little Akt activity could be detected in fibroblasts acutely infected with Ad-␤-Gal (data not shown). In contrast to these results, minimal activation of the MAP kinases, Erks 1 and 2 or p38, was seen in Ad-MyoD or Ad-␤-Gal-infected cells during the same interval (data not shown).
The significance of the rise in Akt kinase activity in MyoD-mediated muscle differentiation was tested by co-infecting 10T1/2 cells with recombinant adenoviruses expressing MyoD and a dominant-negative Akt under control of a tetracyclineregulated gene promoter (Ad-Akt DN ). In pilot studies, Ad-Akt DN blocked IGF-induced Akt enzymatic activity (data not shown). As seen in Fig. 2, A and B, inhibition of endogenous Akt blocked MHC and troponin-T expression and myotube formation but had no effect on production of myogenin. In contrast, when expression of Akt DN was impeded by the tetracycline analog, doxycycline, differentiation proceeded normally. Thus, Akt activity is required for MyoD-stimulated muscle differentiation.

IGF-II mRNA and Protein Expression Are Induced, and the IGF-I Receptor Is Activated during MyoD-mediated Muscle
Differentiation-We next investigated mechanisms responsible for activation of Akt during MyoD-stimulated myoblast differentiation. We first examined components of the IGF system, as IGF signaling has been found to stimulate Akt and to enhance differentiation of established muscle cell lines and in vivo (22,28). As shown in Fig. 3A, IGF-II gene expression, measured by semi-quantitative RT-PCR assay, was progressively induced in Ad-MyoD infected fibroblasts beginning by 8 h after incubation in DM. In contrast, little IGF-I mRNA could be detected over the same time frame, and transcripts for IGF-II and myogenin were not seen in fibroblasts infected with Ad-␤-Gal and incubated in DM for the same time period. Transcripts for ribosomal protein S17 did not change in abundance.
IGF-II protein expression also was induced in Ad-MyoDinfected fibroblasts. As assessed by immunoblotting of condi-tioned DM, IGF-II accumulation was observed in concentrated media from cells infected with Ad-MyoD after 1 day and showed a dramatic increase by 2 days (Fig. 3B). Little IGF-II could be seen by 12 h in DM (data not shown), and none was detected in concentrated conditioned media from fibroblasts infected with Ad-␤-Gal. Thus, induction of IGF-II gene and protein expression are early events in MyoD-regulated myoblast differentiation.
IGF-II activates intracellular signal transduction pathways by binding to the IGF-IR, a ligand-stimulated tyrosine protein kinase that undergoes autophosphorylation as an initial event in its activation (20). To assess phosphorylation of the IGF-IR in Ad-MyoD-infected 10T1/2 fibroblasts, lysates from cells incubated in DM were immunoprecipitated with an antibody to the ␤ subunit of the receptor, followed by immunoblotting with an antibody to phosphotyrosine. As shown in Fig. 3C, progressively increasing tyrosine phosphorylation of the IGF-IR was detected beginning at 1 day after incubation of cells in DM. By 2 days, the extent of receptor tyrosine phosphorylation exceeded that induced in 10T1/2 cells after incubation with IGF-I for 15 min. By contrast, little receptor phosphorylation was seen after 12 h in DM in Ad-MyoD-infected fibroblasts (data not shown) or in Ad-␤-Gal-infected 10T1/2 cells after up to 2 days in DM (Fig. 3C).
IGF-II Action Is Required for MyoD-stimulated Muscle Differentiation-Experiments were performed next to assess the functional significance of production of IGF-II and activation of the IGF-IR for MyoD-mediated myoblast differentiation. Fibroblasts were co-infected with Ad-MyoD and a recombinant adenovirus encoding an antisense cDNA for mouse IGF-II under control of a tetracycline-inhibited promoter (Ad-IGF-II AS ). As shown in Fig. 4A, Ad-IGF-II AS caused a marked decline in induction of IGF-II mRNA after a 1 day incubation of cells in DM, which was reversed by doxycycline. Secretion of IGF-II also was blocked in cells expressing IGF-II AS mRNA, and tyrosine phosphorylation of the IGF-IR was impaired, being seen only at the 2-day time point (Fig. 4, B and C). This apparent discrepancy may be explained by the partial inhibition of Magnification is ϫ200. B, immunoblots for myogenin, MHC, phospho-Akt (pAkt S473 ), and Akt using whole cell protein lysates from 10T1/2 fibroblasts infected with Ad-MyoD or Ad-␤-Gal for the times indicated. C, results of enzymatic assay for Akt activity using immunoprecipitates from Ad-MyoD infected 10T1/2 fibroblasts incubated in DM for the times indicated, and GSK-3␤ as substrate. "Con" and "IGF-I" represent results from uninfected 10T1/2 fibroblasts incubated in DM without or with IGF-I (2 nM of R3-IGF-I analog) for 15 min, respectively. Comparable results were seen for A-C in five independent experiments.

FIG. 2. Reversible inhibition of MyoD-induced muscle differentiation by forced expression of a recombinant adenovirus encoding dominant-negative Akt.
Results are shown of time course experiments using 10T1/2 cells infected with Ad-MyoD and Ad-Akt DN with or without doxcycline (Dox (500 ng/ml)) and incubated in DM for up to 2 days. A, immunocytochemistry for MHC (red) and myogenin (green). Magnification is ϫ200. B, immunoblots for myogenin, MHC, troponin-T, and ␣-tubulin using whole cell protein lysates. Analogous results were observed for A and B in three independent experiments. IGF-II mRNA by Ad-IGF-II AS and by the greater sensitivity of the assay for IGF-IR tyrosine phosphorylation than the assay for IGF-II protein (with a detection limit of ϳ3 nM of IGF-II). Both were restored to normal with doxycycline (Fig. 4, B and  C). Thus, Ad-IGF-II AS impaired both IGF-II production and IGF-IR activation.
Inhibition of IGF-II also diminished MyoD-mediated differentiation. As shown in Fig. 4, D and E, Ad-IGF-II AS reduced the rise in myogenin accumulation, blocked MHC expression, prevented myotube formation, and delayed Akt phosphorylation. In contrast, when expression of IGF-II AS mRNA was blocked by doxycycline, Akt phosphorylation was restored and differentiation proceeded normally. Thus, early production of IGF-II and activation of the IGF-IR are required for MyoD-stimulated muscle differentiation.
The central role of MyoD and related bHLH transcription factors in muscle cell specification and differentiation has been known for over a decade (3), and numerous studies have demonstrated that MyoD can readily convert a range of cell types to myoblasts (3). We now find that an endogenously initiated signaling pathway, involving induction of IGF-II gene and protein expression, and stimulation of the IGF-IR and Akt, are additional key components of MyoD-mediated myoblast differentiation. IGF-II production, and IGF-IR and Akt activation, are relatively early events in the actions of MyoD in this model system, occurring soon after induction of myogenin, and this signaling pathway appears necessary for differentiation to proceed, as either inhibition of IGF-II or blockade of Akt impaired expression of MHC and troponin-T and formation of multinucleated myofibers. Thus, at least in the context of this model, our results indicate that IGF-II functions to initiate an essential autocrine amplification cascade for MyoD-mediated differentiation. Our observations also identify an approach that will be useful in defining the critical signaling pathways that act downstream of Akt in muscle cells.
The biochemical mechanisms by which MyoD induces IGF-II gene expression are unknown. Previous studies have estab-lished that IGF-II gene transcription is stimulated during differentiation of established muscle cell lines (29) but have not identified key DNA response elements or defined critical transcription factors. The mouse IGF-II gene is complicated. It contains three tandem promoters, each with unique 5Ј noncoding exons (30). IGF-II gene expression is also regulated by genomic imprinting, and the gene resides ϳ70 kb 3Ј to the H19 gene within a large imprinted locus on mouse chromosome 7 (31). The two genes are reciprocally imprinted with IGF-II being expressed from the paternally derived chromosome and H19 from the maternal chromosome (31). Investigation into the mechanisms of imprinting has led to identification of a key paternally methylated genomic region located just 5Ј to H19 that functions as an insulator element and regulates reciprocal expression of the two genes (32,33). It has been postulated that binding of the nuclear zinc finger protein CTCF to the unmethylated (maternal) chromosome at this site directs activity of an enhancer located 3Ј to the H19 locus to the nearby H19 promoter, while lack of binding of CTCF to the methylated (paternal) chromosome directs enhancer function to the further Results are shown of time course experiments using 10T1/2 fibroblasts infected with Ad-MyoD and a recombinant adenovirus encoding IGF-II in the antisense orientation (Ad-IGF-II AS ) and incubated in DM in the presence or absence of the tetracycline analog, doxycycline (Dox) for up to 2 days. A, reversible inhibition of IGF-II gene expression by Ad-IGF-II AS , as measured by semiquantitative RT-PCR. Doxycycline prevents expression of IGF-II AS mRNA. B, reversible inhibition of IGF-II protein accumulation. C, reversible impairment of tyrosine phosphorylation of the IGF-I receptor ␤ subunit (IGF-IR␤) by IGF-II AS mRNA, as assessed after immunoprecipitation (IP) with anti-phosphotyrosine (pTyr) and immunoblotting (IB) with anti-IGF-IR␤. D, reversible inhibition of MHC expression and myotube formation by IGF-II AS mRNA. Immunocytochemistry is shown for MHC (red) and myogenin (green). Magnification is ϫ200. E, reversible impairment of myogenin, troponin T, and MHC expression, and Akt phosphorylation (pAkt S473 ) demonstrated by immunoblotting. MyoD protein expression is constant. Comparable results were observed for A-E in three independent experiments. 5Ј IGF-II gene (32,33). It is now possible to determine whether this region or other nearby sites containing putative enhancer elements (34 -36) are critical for MyoD-induced IGF-II gene activity in muscle.