Selective Control of Skeletal Muscle Differentiation by Akt1*

The phosphatidylinositol 3-kinase-Akt pathway plays a central role in growth, development, and metabolism in both normal and neoplastic cells. In skeletal muscle, Akt has been implicated in regulating regeneration and hypertrophy and in counteracting atrophy. Here we provide evidence that Akt1 and not Akt2 is essential for muscle differentiation. Using a robust model of MyoD-mediated muscle development, in which dominant-negative Akt blocked differentiation, we show that targeted loss of Akt1 was equally inhibitory. Selective elimination of Akt1 had no effect on myoblast viability or proliferation but prevented differentiation by impairing the transcriptional actions of MyoD. In contrast, knockdown of Akt2 had no effect on myoblast survival or differentiation and minimally inhibited MyoD-regulated transcription. Our results define isoform-specific Akt-regulated signaling pathways in muscle cells that act through Akt1 to sustain muscle gene activation and promote differentiation.

The phosphatidylinositol 3-kinase-Akt pathway plays a central role in growth, development, and metabolism in both normal and neoplastic cells. In skeletal muscle, Akt has been implicated in regulating regeneration and hypertrophy and in counteracting atrophy. Here we provide evidence that Akt1 and not Akt2 is essential for muscle differentiation. Using a robust model of MyoD-mediated muscle development, in which dominant-negative Akt blocked differentiation, we show that targeted loss of Akt1 was equally inhibitory. Selective elimination of Akt1 had no effect on myoblast viability or proliferation but prevented differentiation by impairing the transcriptional actions of MyoD. In contrast, knockdown of Akt2 had no effect on myoblast survival or differentiation and minimally inhibited MyoD-regulated transcription. Our results define isoform-specific Akt-regulated signaling pathways in muscle cells that act through Akt1 to sustain muscle gene activation and promote differentiation.
The serine-threonine protein kinases of the Akt family have been subjects of intense scrutiny since their discovery in the early 1990s (1,2). The three Akts share a common structure and ϳ85% amino acid identity and are activated by growth factors and hormones through similar biochemical mechanisms that are initiated by production of the lipid signaling molecule, phosphatidylinositol-3,4,5 trisphosphate (1,2). Initial studies of the effects of these proteins established roles for them in cell proliferation, survival, and metabolism (1,2), whereas more recent analyses in mice revealed distinct developmental functions for each Akt. Akt1 deficiency led to a reduction in somatic growth (3); targeted loss of Akt2 caused insulin resistance and diabetes mellitus and growth impairment (4); and knock-out of Akt3 led to reduced brain size (5). Combined deficiency of Akt1 and Akt2 was associated with severe growth defects and neonatal death (6), whereas targeted loss of Akt1 and Akt3 was lethal in mid-gestation (7).
Skeletal muscle represents a tissue that is responsive to Akt, and Akt action has been linked to skeletal muscle development, regeneration, and hypertrophy through several pathways that culminate in stimulation of protein synthesis, inhibition of atrophy, and prevention of cell death (8,9). In mice, both Akt1 and Akt2 appear to be important for muscle development as their combined deficiency resulted in severe hypoplasia (6) but also caused defects in other tissues and organs. Despite these observations, it has not been established which Akt might maintain myoblast survival, promote regeneration, or induce hypertrophy. In cultured muscle cells, Akt2 is induced during differentiation, whereas Akt1 is constant (10 -12). A constitutively active Akt can stimulate muscle hypertrophy in transgenic mice and tissue culture (13,14) and can inhibit atrophy induced by denervation (15), but it is likely that this overexpressed protein does not faithfully mimic physiologically relevant signaling pathways. Here we have assessed the roles of Akt1 and Akt2 in muscle. We find through selective gene knockdown that Akt1 is necessary for initiation and maintenance of myoblast differentiation but that Akt2 is dispensable. Reduction of each Akt led to similar alterations in phosphorylation of several Akt substrates, but only loss of Akt1 inhibited activity of the myogenic transcription factor MyoD. Our results define functionally separable Akt-mediated signaling mechanisms that act primarily through Akt1 to promote and sustain muscle differentiation.
Construction and Use of Recombinant Adenoviruses-The following have been described: mouse MyoD (Ad-MyoD), dominant-negative Akt (Ad-Akt DN ), mouse IGF-II in the antisense orientation (Ad-IGF-II AS ), tetracycline-inhibited transcriptional activator (Ad-tTA), ␤-galactosidase (Ad-␤-Gal) (16,17). Adenoviruses expressing short hairpin inhibitor RNAs targeting mouse Akt1 or Akt2 mRNAs were prepared after testing three sets of double-stranded oligoribonucleotides for each Akt. Akt-selective hairpin oligonucleotides plus a RNA polymerase III termination signal were cloned into an adenoviral shuttle plasmid 3Ј to the human H1 RNA promoter (18). The sense strand is as follows: mouse Akt1, 5Ј-CTAGTGTGAGG-TTGACAGAGGAACTTCAAGAGAGTTCCTCTGTCAAC-CTCACTTTTTGGATCCA-3Ј; mouse Akt2, 5Ј-CTAGTGCC-AACCTTGGCTGTTACATTCAAGAGATGTAACAGCCA-AGGTTGGCTTTTTGGATCCA-3Ј. Akt-specific sequences are underlined. Adenoviruses were generated, purified on discontinuous cesium chloride gradients, and titered by optical density (16). * These studies were supported by National Institutes of Health Grant RO1 DK42748 to (P. R.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Infections with Ad-MyoD, Ad-IGF-II AS , and Ad-Akt DN were performed as described (19). For infections with Ad-␤-gal, Ad-shAkt1, or Ad-shAkt2, viruses were added to cells at ϳ25% of confluent density (multiplicities of infection of 500) followed 24 h later by infection with Ad-MyoD (multiplicity of infection of 125). After an additional 24 h in growth medium, cells were washed, and DM was added.
Immunocytochemistry-All steps have been described (16), as have calculation of the cell number, fusion index, and myotube area (17).
Luciferase Reporter Gene Assays-Reporter plasmids and luciferase assays have been described (19). Cells were seeded at 1 ϫ 10 5 /12-well dish, and the next day, they were transfected with 0.2 g of plasmid DNA using TransIT LT-1 (Mirus Corp, Madison, WI). Cells were infected sequentially 16 h later with Ad-␤-gal, Ad-shAkt1, or Ad-shAkt2 or both Ad-shAkt1 and Ad-shAkt2 followed by Ad-MyoD, as above.
Statistical Analysis-Data are presented as mean Ϯ S.D. Statistical significance was determined by paired Student's t test (p Ͻ 0.05).

Inhibition of Akt Activity Blocks MyoD-mediated Muscle
Differentiation-We previously defined an autocrine growth factor circuit involving IGF-II, the IGF-I receptor, and the PI3-kinase-Akt pathway that was necessary for differentiation of muscle cell lines (16,17) and mesenchymal stem cells acutely converted to myoblasts by MyoD (16). The importance of Akt was assessed in Fig. 1, which shows that muscle differentiation can be reversibly blocked by inhibition of IGF-II production or by dominant-negative Akt (Akt DN ). Adenoviruses encoding IGF-II AS or Akt DN not only led to loss of induction of Akt phosphorylation at serine 473 but also prevented accumulation of muscle proteins, including myogenin and troponin-T, yet had no effect on adenoviralderived MyoD (Fig. 1A). Akt DN blocked the normal increase in the enzymatic activity of Akt (Fig. 1B) and completely inhibited morphological differentiation (Fig. 1C), as was seen in myoblasts lacking IGF-II (16,19).
To address mechanisms by which Akt drives muscle differentiation, we assessed the effects of Akt DN on transcriptional actions of MyoD, using a promoter-reporter gene containing four copies of a MyoD-responsive E-box. We found recently that stimulation of the IGF-I receptor by myoblast-derived IGF-II was required for sustained MyoD-mediated muscle gene activation (19). E-box-dependent promoter activity was reduced by 85% by Akt DN or IGF-II AS but increased to control values when doxycycline inhibited production of these molecules (Fig. 1D). Thus, interference with Akt-mediated signaling blocks muscle differentiation by impairing transcriptional actions of MyoD.

Akt1 Is Critical for Initiation and Maintenance of Muscle
Differentiation-To address the roles of individual Akt isoforms in differentiation, we developed adenoviruses encoding short hairpin interfering RNAs targeting Akt1 or Akt2 mRNAs. Ad-shAkt1 specifically reduced levels of Akt1 mRNA in MyoDexpressing stem cells by 90% but had no effect on transcripts for Akt2 or S17, and Ad-shAkt2 specifically decreased Akt2 mRNA levels ( Fig. 2A). When cells were infected with Ad-shAkt1 and Ad-shAkt2, transcripts for both Akts were eliminated ( Fig. 2A).
Knockdown of Akt1 mRNA caused a 90% reduction of Akt1 protein levels but did not prevent the normal increase in Akt2 seen during differentiation (10 -12) and did not change the amount of adenoviral-derived MyoD (Fig. 2B). Deficiency of Akt1 led to a marked reduction in myogenin and troponin-T (Fig. 2B) and caused an 80% decrease in myotube formation (Fig. 2, C and D) and a 90% decline in myocyte fusion index when compared with controls (Fig. 2E). It is likely that this residual muscle differentiation reflects the ϳ15% of cells that were not infected with Ad-shAkt1. Knockdown of Akt2 mRNA caused the near disappearance of Akt2 protein but had little effect on Akt1 or MyoD and did not inhibit induction of myogenin or troponin-T (Fig. 2B). Loss of Akt2 also did not reduce myofiber formation or fusion index (Fig. 2, C-E). The combination of Ad-shAkt1 and Ad-shAkt2 led to the near elimination of both Akt proteins and blocked differentiation to the same extent as Ad-shAkt1 alone (Fig. 2B). We observed identical results in the C2 muscle cell line, where Akt1 deficiency inhibited differentiation at its earliest stages and knockdown of Akt2 did not (supplemental Fig. 1).
Deficiency of either Akt1 or Akt2 did not diminish cell proliferation during incubation in growth medium and did not increase myoblast death during 48 h in DM (Fig. 2F). Only loss of both Akts led to reduction in total cell number after 24 or 48 h in DM (17-20%; Fig. 2F). Overall, the minimal cell death probably reflects the effectiveness of MyoD in sustaining myoblast survival when overexpressed (20), although Akt has been shown to be a key agent in maintaining muscle cell viability (20,21).
Assessing Signaling Pathways Activated by Akt1 and Akt2 in Muscle Cells-To address mechanisms by which Akt1 and Akt2 differentially influenced differentiation, we examined the effects of reduction of each protein on selected Akt target proteins. Akt1 deficiency led to a greater decline in total Akt and phosphorylation at serine 473 than did loss of Akt2 (Fig. 3A). Deficiency of Akt1 or Akt2 led to reductions in inhibitory phosphorylation of GSK3␣ and ␤ on serine residues 21 and 9. Akt1 loss caused a greater decline in abundance of p70S6 kinase and in its activating phosphorylation by mTOR than did loss of Akt2. Reduction of Akt1 led to a slight rise in inhibitory phosphorylation of FoxO3a (23), although neither Akt appeared to modify the decline in abundance of FoxO3a seen during incubation in DM. Similar alterations in Akt target proteins were observed in C2 cells (data not shown). Thus, loss of each Akt appears to lead to analogous changes in some signaling proteins and different effects on others.
Lack of Akt1 Impairs Transcriptional Functions of MyoD-We next asked whether Akt deficiency inhibited MyoD-mediated transcription. Myogenin promoter activity was reduced by 85% in cells without Akt1, whereas loss of Akt2 caused a 40% increase. Muscle creatine kinase promoter activity also fell by ϳ90% in cells lacking Akt1, whereas Akt2 deficiency caused a 50% decline (Fig. 3B). Loss of both Akts yielded results similar to lack of Akt1 (Fig. 3B). Thus, Akt1 and Akt2 exert distinct effects on muscle gene activation. Akt1 appears to be needed for induction of muscle gene expression early in differentiation, thus explaining why its loss prevents differentiation. By contrast, Akt2 is dispensable for early muscle gene expression but may contribute to regulation of later genes.
Akts in Cell Fate and Function-Our results add to a perplexing literature on specific actions of individual Akts in different cell lineages. For example, viability of suprabasal keratinocytes and their subsequent differentiation was reduced by loss of Akt1 but not by lack of Akt2 and could be restored to normal by preventing cell death with a broad spectrum caspase inhibitor (24). We find that the same inhibitor did not reverse the lack of muscle differentiation seen with Akt1 deficiency (data not shown), indicating that different mechanisms operate in keratinocytes and myoblasts. By contrast, osteoclast differentiation appears to require Akt1 and Akt2 as knockdown of either protein was inhibitory, without reducing cell viability or proliferation (25). In mature adipocytes or myocytes, Akt2 is essential for normal glucose metabolism, whereas Akt1 was dispensable (26,27). In breast cancer cells, Akt1 appears to control motility and invasiveness, with overexpression being inhibitory and its loss facilitating, as long as Akt2 also was expressed (28). The mechanisms by which Akt1 negatively regulates motility in these cells are not defined. In one study, Akt1 deficiency led to increased Erk activity, which was required for enhanced cell migration (28), whereas in another, Akt1 stimulated degradation of the transcription factor NFAT1 by the E3 ubiquitinprotein isopeptide ligase, MDM2 (29). Thus, depending on the cell or tissue type, Akt1 or Akt2 exert either similar or distinct effects on differentiation, survival, intermediary metabolism, or motility. The steps by which individual Akts couple selectively with specific downstream effector pathways in different cell types to yield unique outcomes remain an enigma.
Akts and Skeletal Muscle-A growing literature supports roles for Akt proteins in muscle growth and regeneration after injury (8,9). Akts inhibit FoxO transcription factors by direct phosphorylation, thereby blocking expression of FoxO-activated genes (23). In muscle, this prevents induction of atrogin-1/MAFbx and MuRF1, E3 ubiquitin ligases involved in promoting atrophy (30,31). In transgenic mice, overexpression of active Akt led to extensive hypertrophy (14), as did injection of individual mouse or rat muscles with a similar expression vector (13,15). Overexpression of active Akt in rodents also reduced muscle atrophy after denervation and enhanced repair after injury, effects blocked by the mTor inhibitor, rapamycin (13,15), implying that selective pathways downstream of Akt are critical to regeneration. Myofiber hypertrophy was seen in transgenic mice overexpressing IGF-I in muscle (32), and muscle damage was diminished in mice with muscular dystrophy by an IGF-I transgene, coincident with sustained Akt phosphorylation (32). Thus, Akts mediate many actions of IGF-I in muscle, but the specific responsible Akt remains an open question.
We recently established that autocrine signaling mediated by the IGF-I receptor, PI3-kinase, and Akt, was necessary for sustained transcriptional actions of MyoD shortly after onset of muscle differentiation (19). Inhibition of this pathway blocked FIGURE 3. Alterations in signaling proteins and inhibition of MyoD-mediated transcription in Akt-deficient muscle cells. A, results are shown of 10T1/2 cells infected with Ad-MyoD plus either Ad-shAkt1 or Ad-shAkt2 followed by incubation in DM for 0, 24, or 48 h. Immunoblots of protein lysates are seen for pAkt Ser473 , total Akt, phospho-GSK3␣ serine 21 and phospho-GSK3␤ serine 9 (pGSK3␣ S21 /␤ S9 ), pGSK3␤ S9 (short exposure), GSK3␤, phospho-p70S6 kinase T389 (p-p70S6K T389 ), p70S6K, phospho-FoxO3a (pFoxO3a), FoxO3a, and ␣-tubulin. ns, nonspecific bands. B, promoter-reporter gene experiments were performed in 10T1/2 cells infected with Ad-MyoD plus either Ad-shAkt1, Ad-shAkt2, or both viruses followed by incubation in DM for 24 h. The left panel shows results with the proximal mouse myogenin promoter (mean Ϯ S.D. of three experiments: *, p Ͻ 0.00002, **, p Ͻ 0.002, ***, p Ͻ 0.002 versus control). The right panel depicts results with the mouse muscle creatine kinase promoter (mean Ϯ S.D. of three experiments: *, p Ͻ 0.0006, **, p Ͻ 0.003, ***, p Ͻ 0.005 versus control). Luc, luciferase. C, model of regulation of muscle differentiation by Akt1 as part of an IGF-II-initiated autocrine amplification pathway. IGF-II, produced by differentiating myoblasts, activates the IGF-I receptor (IGF-IR), adaptor proteins IRS1 and 2, and the PI3-kinase-Akt pathway, including both Akt1 and Akt2. Akt1 then enhances transcriptional actions of MyoD via co-activators p300 and P/CAF. Except for down-regulation of Akt2, all other indicated inhibitory steps block muscle differentiation. Unknown outcomes are depicted by a question mark. si-RNA, small interfering RNA. LY, LY294002. ACCELERATED PUBLICATION: Akt1 and Muscle Differentiation FEBRUARY 23, 2007 • VOLUME 282 • NUMBER 8 myogenin gene transcription without altering MyoD or its DNA binding partner E12/E47. Rather, inhibition of IGF signaling reduced accumulation of transcriptional co-activators p300 and P/CAF (p300/CBP-associated factor) on chromatin at the myogenin promoter, thereby preventing histone acetylation and impairing recruitment and activation of RNA polymerase II (19). Taken together with current observations, these results implicate Akt1 as a key intermediate in MyoD-mediated transcriptional pathways in muscle cells, although the mechanisms by which Akt1 sustains activity of MyoD remain to be elucidated. A model outlining our findings is shown in Fig. 3C.
Previous observations indicating that Akt2 increases in abundance during muscle differentiation (10 -12) and that its neutralization can inhibit differentiation (10) have led to the idea that Akt2 rather than Akt1 plays a central role in muscle. In contrast, our results demonstrate clearly that Akt1 is critical for induction and maintenance of muscle differentiation, although we did note subtle deficits in Akt2-deficient myotubes. In MyoD-converted 10T1/2 cells, lack of Akt2 led to reduction in overall myofiber length without a decrease in fusion index (Fig.  2, C and E), whereas in C2 cells, its loss was associated with thinner myotubes (supplemental Fig. 1). Two molecules that act downstream of Akt have been linked to myofiber formation, although the mechanisms are unknown. The protein kinase mTor regulates protein synthesis and is activated indirectly by Akt (22). C2C12 myocytes overexpressing an enzymatically inert mTor developed smaller myotubes than did cells with wild-type protein (33). The transcription factor FoxO1 was found to accelerate fusion and myotube formation when overexpressed (34), although other results showed that FoxO1 inhibited differentiation (35).
Active signaling through the IGF-I receptor has been associated with increased cancer risk and progression (36) and accelerated aging in experimental animals (37). At the same time, treatment with IGF-I has been proposed for muscle disorders and to counteract sarcopenia in aging humans (8,9). Optimal therapeutic use of this growth factor in muscle or other tissues will require a fuller understanding of its key signaling molecules and their mechanisms of action.