IGF-II transcription in skeletal myogenesis is controlled by mTOR and nutrients

The insulin-like growth factors (IGF-I and IGF-II) are critically involved in skeletal muscle development (Florini et al., 1991a) as well as adult muscle regeneration and hypertrophy (Barton-Davis et al., 1999), which require muscle satellite cell activation and differentiation (Rosenblatt et al., 1994). The autocrine actions of IGF-II are essential for the differentiation of satellite cells in culture (Florini et al., 1991c). However, the regulation of IGF-II expression in these cells is poorly understood. The IGF-II gene is complex in structure and its expression can be regulated at many levels, including transcription from multiple promoters, alternative translation initiation, mRNA stability, and genomic imprinting. IGF-II expression during skeletal muscle differentiation is regulated at the transcriptional level, and the transcripts are derived from promoter 3 (P3; Kou and Rotwein, 1993). However, the activity of this promoter was not increased in myotubes compared with myoblasts, and further analyses of a 25-kb segment around this promoter region did not reveal any element responsible for myogenic regulation of IGF-II transcription (Kou and Rotwein, 1993). Partly due to the lack of information on the relevant cis-acting regulatory elements, the upstream pathways that directly instruct the expression of IGF-II in muscle differentiation are unexplored.

The cellular target of the bacterial macrolide rapamycin, mTOR, belongs to the phosphatidylinositol kinase (PI3K)-related family of Ser/Thr kinases and functions as a master regulator of cell growth and proliferation by regulating multiple downstream effectors (Jacinto and Hall, 2003). In cell proliferation, the best-characterized function of mTOR is the regulation of translation initiation through eIF-4E binding protein 1 (4EBP1) and S6 kinase 1 (S6K1), and the PI3K pathway acts in parallel with mTOR to regulate 4EBP1 and S6K1 (Fumagalli and Thomas, 2000; Gingras et al., 2001). The mTOR pathway is believed to mediate nutrient signals such as amino acid sufficiency (Fumagalli and Thomas, 2000; Gingras et al., 2001), as well as directly receive mitogenic signals through a lipid second messenger (Fang et al., 2001).

Implicated by the inhibitory effect of rapamycin on the differentiation of a variety of myoblasts in culture (Coolican et al., 1997; Cuenda and Cohen, 1999; Conejo et al., 2001; Erbay and Chen, 2001), mTOR's essential role in skeletal myogenesis has been demonstrated by the ability of a rapamycin-resistant mTOR mutant to rescue rapamycin-inhibited differentiation in C2C12 myoblasts (Erbay and Chen, 2001; Shu et al., 2002). Remarkably, the kinase activity of mTOR is not required for initiation of differentiation, and both S6K1 and 4EBP1 have therefore been excluded as downstream mediators of mTOR's myogenic signaling (Erbay and Chen, 2001). The molecular events regulated by mTOR in myogenesis are currently unknown. In this report, we provide strong evidence that mTOR governs myoblast differentiation by controlling the transcription of IGF-II, potentially through a nutrient-sensing pathway.

To facilitate the investigation of IGF-II regulation during myogenesis, we established C2C12 cell pools stably expressing a luciferase reporter for P3 of the IGF-II gene. P3 activity was found slightly decreased in fully formed myotubes when compared with myoblasts (unpublished data), this finding is in agreement with a previous work (Kou and Rotwein, 1993). However, a closer examination revealed an acute activation of P3 3 h after the induction of differentiation by growth factor deprivation, which returned to the basal level by 6 h (Fig. 1 A). As a control, a CMV promoter reporter displayed no change in luciferase activity under the same conditions (unpublished data). Remarkably, this increased activity was completely blocked by rapamycin treatment (Fig. 1 A), whereas the CMV-driven luciferase activity was not affected (not depicted).

This brief activation of P3 alone is insufficient to account for the continued increase of IGF-II mRNA levels throughout differentiation (see Fig. 2 A), which prompted us to consider other potential regulatory elements. A recently identified skeletal ME shared by IGF-II and H19, a gene closely linked to the IGF-II gene and reciprocally imprinted, appeared as a likely candidate (Kaffer et al., 2001). Indeed, the luciferase activity in C2C12 cells stably expressing the ME reporter under the control of H19 promoter increased steadily during differentiation, reaching ∼5-fold by 24 h and ∼40-fold by 48 h when compared with 0 h (Fig. 1 B). The activity of an enhancerless reporter for H19 promoter did not change under similar conditions (unpublished data). Importantly, the ME reporter activity was highly sensitive to rapamycin (Fig. 1 B). Thus, a rapamycin-sensitive pathway may play a distinct role in the regulation of IGF-II transcription through the sequential activation of P3 and ME.

In addition to the mTOR pathway, the PI3K pathway and p38 MAPK are also required for skeletal myocyte differentiation (Coolican et al., 1997; Kaliman et al., 1998; Cuenda and Cohen, 1999). However, both P3 and ME activity during differentiation were insensitive to the PI3K inhibitor wortmannin and the p38 inhibitor SB202190 (Fig. 1, C and D).

Consistent with rapamycin's effect on P3 and ME of IGF-II, the drastic increase of IGF-II mRNA levels during differentiation was significantly blocked by rapamycin treatment of differentiating cells, as shown by the results of RNase protection assays (RPAs) in Fig. 2 A. This inhibition was not due to changes in IGF-II mRNA stability because during a 12-h window of actinomycin D treatment to suppress transcription, IGF-II mRNA remained highly stable, regardless of the presence of rapamycin (Fig. 2 B).

To assess whether the rapamycin effect on IGF-II expression was directly attributable to mTOR, we examined IGF-II mRNA in C2C12 cell lines stably expressing various recombinant FLAG-tagged mTOR proteins including wild type, rapamycin resistant (RR; S2035T), and RR and kinase inactive (RR/KI; S2035T/D2357E; Brown et al., 1995). Several independent clones for each cell line were examined. The expression of IGF-II mRNA in both RR and RR/KI cells, but not in wild-type mTOR cells, was insensitive to rapamycin treatment. Fig. 2 (C–E) shows the RPA results from two representative clones for each cell line. Thus, rapamycin-resistant mTOR, both kinase active and inactive, fully rescued IGF-II expression from rapamycin inhibition. Autophosphorylation assays indicated no detectable mTOR catalytic activity from the recombinant RR/KI proteins in the stable cells (Fig. 2 F). As expected, the endogenous S6K1 activity was abolished by rapamycin treatment in wild-type and RR/KI cells, whereas the RR cells yielded a significant portion of RR S6K1 activity (Fig. 2 G).

Together, our data suggest that during skeletal myogenesis mTOR regulates the production of IGF-II mRNA. Moreover, this myogenic function of mTOR is independent of its kinase activity, which is consistent with our previous observations (Erbay and Chen, 2001). This conclusion contradicts that of Shu et al. (2002) who reported that a C2C12 cell line stably expressing a kinase-inactive mTOR did not differentiate in the presence of rapamycin. It is noted that a different mutation (D2338A) was used to inactivate the kinase by Shu et al. Although D2357E and D2338A are equally effective in inactivating mTOR's catalytic activity (Brown et al., 1995), it cannot be ruled out that D2338A may affect an additional biochemical property of mTOR. The possibility of clonal variation may also be considered. To avoid such a potential problem, multiple stable clones were examined in our studies as described in the previous paragraph, and transient transfection was previously used to express mTOR (Erbay and Chen, 2001).

Because the mTOR pathway is known to sense the availability of amino acids in proliferating cells (Fumagalli and Thomas, 2000; Gingras et al., 2001), we hypothesized that mTOR may mediate amino acid signals in the regulation of IGF-II transcription during myogenesis. Indeed, the acute activation of P3 normally seen at 3 h was abrogated when amino acids were depleted from the differentiation medium (Fig. 3). When normal concentrations of amino acids were replenished for 3 h, activation of P3 was recovered, and this reactivation was completely blocked by rapamycin, but not by wortmannin or SB202190 (Fig. 3). Similarly, the induction of ME activity at 3 h was ablated by amino acid deprivation and stimulated upon readdition of amino acids (Fig. 3). Again, the reactivation of ME by amino acids was inhibited by rapamycin and was insensitive to wortmannin and SB202190. The short duration of amino acid withdrawal ensured that the effect would be the result of affecting a signaling cascade rather than a general stoppage of protein synthesis. Indeed, the CMV-driven luciferase activity was unaffected by amino acid deprivation under similar conditions (unpublished data).

Circulating IGF levels in animals have previously been proposed to provide a critical link between the nutritional status and animal/organ growth (Pilistine et al., 1984), but molecular mechanisms underlying nutritional regulation of IGF levels are unknown. Our results provide the first direct demonstration of IGF-II transcriptional regulation by cellular amino acid levels, and implicate a potential role for mTOR in controlling nutritional regulation of IGF-II expression in skeletal muscle development.

The regulation of IGF-II transcription by mTOR led us to hypothesize that mTOR might exert its myogenic function by controlling IGF-II secretion that is critical for initiation of the differentiation program. To probe this possibility, conditioned medium from normally differentiating “donor” C2C12 cells were transferred daily to “recipient” C2C12 cells subjected to rapamycin treatment. The recipient cells proceeded to terminally differentiate despite the presence of rapamycin, evidenced by the formation of sarcomeric myosin heavy chain (MHC)–positive myotubes (Fig. 4 A, a). Thus, it appears that rapamycin's inhibitory effect on differentiation is through the elimination of a secreted factor. This factor was confirmed to be IGF-II by the observation that preincubation with an anti–IGF-II antibody, but not a control rabbit IgG, completely eliminated the myogenic potency of the donor medium (Fig. 4 A, b and c). Moreover, conditioned media from rapamycin-treated RR-mTOR cells, as well as RR/KI-mTOR cells, conferred rapamycin resistance to the differentiation of the recipient cells (Fig. 4 A, d and e), further validating the requirement of mTOR and a kinase-independent mechanism. The neutralizing IGF-II antibody also abolished the rapamycin-resistant myogenic effects of the media from the recombinant mTOR-expressing cells (unpublished data).

Furthermore, exogenous IGF-II (at 300 ng/ml) restored C2C12 differentiation in the presence of rapamycin as shown by myotube formation (Fig. 4 B) with a differentiation index indistinguishable from that of control cells (Fig. 4 C). Similar rescuing effects were also observed with 150 ng/ml IGF-II (unpublished data), a concentration comparable to that reportedly secreted by C2C12 cells (Tollefsen et al., 1989). Interestingly, 100 ng/ml IGF-I was equally effective (Fig. 4, B and C). The apparently thinner myotubes in the IGF-rescued culture suggests that mTOR may play a role in myotube morphology independent of IGF and the initiation of differentiation. Importantly, the differentiation of mouse primary myoblasts was inhibited by rapamycin, which was also reversed by exogenous IGF-I or IGF-II (Fig. 4, B and C), suggesting that the molecular mechanisms delineated in C2C12 cells are most likely applicable to primary cells. In differentiating C2C12 cells, IGF-II expresses to a much higher level than IGF-I (Tollefsen et al., 1989), and IGF-II antisense oligonucleotides are sufficient to abolish differentiation (Florini et al., 1991c). The collective evidence is consistent with the notion that IGF-II is the primary myogenic factor in these cells, whereas IGF-I can compensate for the lack of IGF-II; both IGFs presumably signal through the IGF-I receptor. Together, it is apparent that the primary function of mTOR in the initiation of skeletal muscle differentiation is controlling IGF-II production and not mediating IGF downstream signaling.

The PI3K–Akt pathway has been shown to be critical for IGF's myogenic signaling (Coolican et al., 1997; Kaliman et al., 1998). We found that expression of a constitutively active (c.a.) Akt reversed the inhibition of C2C12 differentiation by the PI3K inhibitor LY294002 (Fig. 5 A), suggesting that Akt is the main effector for PI3K signaling in myogenesis. Remarkably, c.a.Akt also rescued differentiation from rapamycin inhibition (Fig. 5 A). The results of RPA analysis indicated that IGF-II mRNA levels were significantly inhibited by rapamycin in c.a.Akt-expressing cells (Fig. 5 C), which excluded the possibility of enhanced IGF-II production independent of mTOR in c.a.Akt cells. These results imply that the PI3K–Akt pathway is sufficient to mediate IGF's myogenic signaling, and they further confirm that regulation of IGF production is a primary function for mTOR in myogenesis. It is interesting to note the contrast between the relationship of these two pathways in myogenesis and mitogenesis. In cell proliferation, the mTOR pathway and the PI3K pathway cooperate in a parallel manner to transduce growth factor signals and regulate the same downstream targets such as S6K1 (Fumagalli and Thomas, 2000); whereas in myogenesis, they regulate two sequential processes: production of IGF and subsequent IGF signaling, respectively (Fig. 5 D).

In conclusion, we have shown that the autocrine IGF-II transcription required for skeletal myocyte differentiation is regulated by mTOR and the availability of amino acids. Although the expression of various myogenic markers, including myogenin, p21^CIP, and MHC, is also blocked by rapamycin (Cuenda and Cohen, 1999), they are regulated by IGF signaling (Florini et al., 1991b), and are therefore unlikely to be directly controlled by mTOR. To the best of our knowledge, this is the first identification of an upstream signaling pathway in the regulation of IGF production in myogenesis. Importantly, our findings also suggest a potential role for the mTOR–IGF axis as a molecular bridge between nutritional status and skeletal muscle development.

M2 anti-FLAG and HA.11 anti-HA were obtained from Sigma-Aldrich, anti–IGF-II was obtained from R&D Systems, anti-S6K1 was obtained from Upstate Biotechnology, and MF20 antisarcomeric MHC was obtained from the Developmental Studies Hybridoma Bank at the University of Iowa. IGF-I, IGF-II, gelatin, actinomycin D, and LY294002 were obtained from Sigma-Aldrich. Rapamycin, wortmannin, and SB202190 were acquired from Calbiochem. All other cell culture reagents were obtained from Invitrogen.

The following plasmids were gifts from various laboratories: pCEFL-HA-Akt (wild type) and pCMV6-myristoylated-HA-Akt (c.a.) were gifts from N. Ahmed and J. Blenis (Harvard Medical School, Boston, MA), mouse IGF-II cDNA was a gift from P. Rotwein (Oregon Health and Science University, Portland, OR), P3-luc reporter and basic-luc (promoterless) were gifts from A. Murrell (The Babraham Institute, Cambridge, UK; Murrell et al., 2001), and H19-luc was a gift from K. Pfeifer (National Institute of Child Health and Human Development, Bethesda, MD; Pfeifer et al., 1996). CMV-luc was as previously reported (Kim and Chen, 2000). To assemble the skeletal ME reporter H19-luc-ME, the SpeI–EagI enhancer fragment (+23 kb to +27 kb) was inserted into a SpeI–EagI linker present in the XhoI site 3′ of the luciferase in H19-luc.

C2C12 myoblasts were maintained and differentiated as described previously (Erbay and Chen, 2001). Primary skeletal myoblasts were isolated from 2-wk-old mice, grown, and differentiated as previously reported (Rando and Blau, 1994). Transfections were performed using FUGENE-6 (Roche), and stable clones or pools were selected in 1 mg/ml G418. For amino acid deprivation and readdition, the cells were cultured in Dulbecco's PBS with 4.5 g/liter glucose and 2% dialyzed serum for 3 h and switched to normal differentiation medium for 3 h. For immunocytochemistry, cells were fixed in 3.7% formaldehyde and stained with MHC antibody and FITC–anti–mouse IgG. Microscopy was performed on a microscope (model DMIL; Leica) with CPLAN 10×/0.22 NA lenses. The images were captured using a monochrome charged-couple device camera (model SPOT RT; Diagnostic Instruments), and processed as 8-bit RGB images using Adobe Photoshop 7.0.

Total RNA from C2C12 cells (in 60-mm plates) was isolated using the RNeasy mini kit (QIAGEN), and 4 μg of each RNA was applied to RPA using the RPA III kit (Ambion). Radioactive probes (used at 10⁶ cpm/ml) were generated from linearized IGF-II cDNA construct by in vitro transcription using the MAXIscript kit (Ambion). The samples were run on 5% denaturing polyacrylamide gels and analyzed on a phosphorimager (model Cyclone; Packard Instrument Co.). To confirm loading consistency, 1-μg RNA samples were run on 0.8% agarose-formaldehyde gels, followed by ethidium bromide staining.

C2C12 cells stably expressing CMV-luc, P3-luc, basic-luc, H19-luc, or H19-luc-ME were grown to 100% confluence and induced to differentiate in 2% horse serum. Various drug treatments were performed as described in the figure legends. The cells were lysed at the indicated times, and luciferase assays were performed using the Luciferase Assay Systems kit (Promega).

mTOR autokinase assays and S6K1 kinase assays were performed as described previously (Erbay and Chen, 2001).

t tests were performed for all data comparisons. Unless specifically indicated in figure legends, significant difference was defined by P < 0.05.

We thank Drs. Marni Boppart and Stephen Kaufman for insightful comments on the manuscript and Ms. Ai-Luen Wu for technical assistance.

This work was supported by the National Institutes of Health grants GM58064 and AR48914 (to J. Chen) and predoctoral fellowships from the American Heart Association Midwest Affiliate (to E. Erbay and I.-H. Park).