Nuclear Factor κB-inducing Kinase and IκB Kinase-α Signal Skeletal Muscle Cell Differentiation*

Nuclear factor κB (NF-κB)-inducing kinase (NIK), IκB kinase (IKK)-α and -β, and IκBα are common elements that signal NF-κB activation in response to diverse stimuli. In this study, we analyzed the role of this pathway during insulin-like growth factor II (IGF-II)-induced myoblast differentiation. L6E9 myoblasts differentiated with IGF-II showed an induction of NF-κB DNA-binding activity that correlated in time with the activation of IKKα, IKKβ, and NIK. Moreover, the activation of IKKα, IKKβ, and NIK by IGF-II was dependent on phosphatidylinositol 3-kinase, a key regulator of myogenesis. Adenoviral transduction with the IκBα(S32A/S36A) mutant severely impaired both IGF-II-dependent NF-κB activation and myoblast differentiation, indicating that phosphorylation of IκBα at Ser-32 and Ser-36 is an essential myogenic step. Adenoviral transfer of wild-type or kinase-deficient forms of IKKα or IKKβ revealed that IKKα is required for IGF-II-dependent myoblast differentiation, whereas IKKβ is not essential for this process. Finally, overexpression of kinase-proficient wild-type NIK showed that the activation of NIK is sufficient to generate signals that trigger myogenin expression and multinucleated myotube formation in the absence of IGF-II.

The IGFs 1 are the only known growth factors that are crucial to myogenesis (1). IGF-I and IGF-II switch on the myogenic program through the IGF-I receptor (2), activating the expression of myogenic transcription factors, cell cycle arrest, musclespecific protein expression, and cell fusion to form multinucleated myotubes (3,4). PI3K is an essential second messenger for myogenesis (5)(6)(7)(8)(9). We have recently described a myogenic signaling cascade initiated by IGF-II that leads to biochemical and morphological skeletal muscle cell differentiation and that involves PI3K activation, NF-B activation, and inducible nitric-oxide synthase expression and activation (10). In this report, we further analyze the role of the NF-B-activating signaling cascade in myogenesis. NF-B transcription factors are key mediators of inflammatory responses, immune system functioning, transformation, oncogenesis, and anti-apoptotic signaling (11)(12)(13). NF-B exists in the cytoplasm in an inactive form by virtue of its association with inhibitory proteins termed IB (11)(12)(13)(14)(15). NF-B translocation to the nucleus and activation are most frequently achieved through the signal-induced proteolytic degradation of IB in the cytoplasm. Two kinases, IKK␣ and IKK␤, which are contained in a high-molecularweight multiprotein complex, show inducible IB kinase activity and play a key role in NF-B activation by a variety of stimuli (16 -19). Despite their high sequence similarity, IKK␣ and IKK␤ have different regulatory and functional roles. In mice lacking IKK␤, the activation of NF-B by cytokines is abolished, and mouse embryos die on days 12-13 of gestation due to massive liver apoptosis (20). In contrast, IKK␣ is dispensable for pro-inflammatory responses, but plays an essential role in embryonic development. Mice lacking IKK␣ exhibit defective proliferation and differentiation of epidermal keratinocytes and defective limb and skeletal patterning (21,22). IKK␣ and IKK␤ are themselves phosphorylated and activated by one or more upstream kinases, like NIK, which is a member of the mitogen-activating protein kinase kinase kinase family (23)(24)(25).
We report here that IB␣ phosphorylation at Ser-32 and Ser-36 is required for both IGF-II-dependent NF-B activation and differentiation in L6E9 myoblasts. We show that IKK␣ is involved in IGF-II-dependent multinucleated myotube formation and muscle-specific gene expression, whereas IKK␤ is not essential for these processes. Our data suggest that NIK activation triggers myogenin expression and multinucleated myotube formation in the absence of IGF-II.
Cell Culture-Rat L6E9 myoblasts were cultured as described previously (5). Subconfluent myoblasts were differentiated by serum depletion in DMEM plus antibiotics with or without IGF-II (40 nM) in the absence or presence of other compounds, as indicated for each experiment. Cells were photographed after staining the nuclei with Mayer's hemalum solution for microscopy (Merck, Darmstadt, Germany), and cell fusion was quantified by counting nuclei in myotubes from a total of at least 1000 nuclei from 10 -20 randomly selected microscope fields for each condition. For adenoviral transduction, subconfluent L6E9 myoblasts were transduced at a multiplicity of 50 -100 particles/cell and then cultured for an additional 36 h before inducing differentiation with or without IGF-II. To compare the impact of IKK␣ and IKK␤ on myoblast differentiation, experimental conditions were selected to ensure similar levels of expression of IKK␣ and IKK␤ constructs (data not shown).
Construction of Recombinant Adenoviruses-Recombinant adenoviruses expressing FLAG-tagged versions of either wild-type or dominant-negative mutant K44A human IKK␣ and IKK␤ were generated by homologous recombination as described by Graham and Prevec (38). cDNAs were cloned into the shuttle plasmid pAdl1/RSV and cotransfected with pJM17 into 293 cells to achieve homologous recombination. Individual plaques were isolated and checked for recombinant protein expression after infection of 293 cells. Recombinant adenoviruses were further amplified in 293 cells; purified by cesium chloride gradient centrifugation; dialyzed against 1 mM MgCl 2 , 10 mM Tris (pH 7.4), and 10% glycerol; and stored at Ϫ80°C (39). Viral stocks were titrated by infecting 293 cells with serial dilutions of the preparation and observing the cytopathic effect on the cells 48 h after infection. An infectious titer was given assuming that a multiplicity of infection of 10 is required to cause a complete cytopathic effect at 48 h. In addition, the A 260 of the preparation was measured to estimate the particle titer (1 A 260 unit ϭ 10 12 particles/ml).
For immunoprecipitation, antibodies were preadsorbed on protein G-Sepharose at 4°C for 1 h and washed twice in hypotonic solution and 1% Nonidet P-40 before being incubated with the protein extracts for 2 h at 4°C. The immunopellets were washed six times in the same buffer and once in kinase buffer (20 mM Hepes, 10 mM MgCl 2 , 100 M Na 3 VO 4, 20 mM ␤-glycerophosphate, 2 mM dithiothreitol, and 50 mM NaCl (pH 7.5)). Kinase reactions were carried out for 30 min at 30°C using 5 Ci of [␥-32 P]ATP and GST-IB␣-(1-54) as substrate (except for NIK kinase assays, in which autophosphorylation of immunoprecipitated NIK was analyzed). The reaction products were analyzed on 10% polyacrylamide gels and revealed by autoradiography.
Samples (70 g) were incubated for 10 min at 4°C with 30 ng of poly(dI⅐dC) and 5 l of 5ϫ reaction buffer (50 mM Tris (pH 7.5), 500 mM NaCl, 5 mM dithiothreitol, 5 mM EDTA, 20% glycerol, and 0.4 mg/ml salmon sperm DNA) in a final volume of 25 l. The end-labeled probe was added for a further incubation of 25 min at 25°C. The specificity of the bands detected was verified by adding a 10 -100-fold excess of competing unlabeled NF-B probe. NF-B-unrelated oligonucleotide probe controls did not show any specific binding activity (data not shown).
Immunofluorescence-Cells grown on coverslips were fixed for 20 min with 3% paraformaldehyde in PBS, washed three times in PBS, and then treated as follows: (a) 10 min in PBS containing 50 mM NH 4 Cl, (b) 10 min in PBS containing 20 mM glycine, and (c) 30 min in PBS containing 10% fetal bovine serum. Subsequently, coverslips were incubated with primary antibodies (anti-NIK polyclonal antibodies, 1 g/ml; anti-myogenin monoclonal antibody, undiluted culture supernatant) for 1 h at room temperature. After washing in PBS, coverslips were incubated with fluorochrom-conjugated antibodies (Oregon Green or Texas Red) for 45 min. Cells were washed three times in PBS and then mounted in immunofluor medium (ICN Biomedicals Inc., Aurora, OH). Images were obtained using a Leica TCS 4D laser confocal fluorescence microscope with a 40ϫ objective.

IGF-II-induced Skeletal Muscle Cell Differentiation Involves NF-B Signaling Cascade Activation-IGF-II induces NF-B
DNA-binding activity as an early event during L6E9 myoblast differentiation (10). Most extracellular stimuli that activate the NF-B pathway induce phosphorylation of the NF-B repressor IB␣ at Ser-32 and Ser-36 as a requisite for its degradation (19). However, alternative pathways for NF-B activation have been described (26,27). To analyze the mechanism required by IGF-II for NF-B activation during differentiation, L6E9 myoblasts were transduced with an adenovirus expressing an IB␣ mutant with Ser-32 and Ser-36 replaced by alanine residues (adv/IB␣(S32A/S36A)). This mutant inhibits NF-B in pathways involving serine phosphorylation and subsequent proteasome degradation of IB␣. L6E9 myoblasts differentiated for 24 h with IGF-II exhibited an induction of NF-B DNA-binding activity (Fig. 1A), which was blocked in myoblasts overexpressing IB␣(S32A/S36A) (Fig. 1B). Myoblasts infected with an adenovirus expressing green fluorescent protein (adv/GFP) were used as control (Fig. 1B). Next, we analyzed the impact of IB␣ phosphorylation on IGF-II-dependent myoblast differentiation. After 4 days of IGF-II treatment, the expression of muscle-specific proteins such as myosin heavy chain and caveolin-3 was highly decreased in myoblasts overexpressing the IB␣(S32A/S36A) mutant (50 Ϯ 12%, n ϭ 3) compared with non-transduced control cells or cells transduced with adv/GFP, whereas the expression of the non-muscle-specific protein ␤-actin was similar under all conditions (Fig. 1C). Moreover, cells overexpressing the IB␣(S32A/S36A) mutant did not fuse to myotubes, whereas control cells (adv/GFP) showed 82% of the nuclei in myotubes from a total of 1661 nuclei randomly counted (Fig. 1D). These data suggest that IGF-II requires NF-B activation through a mechanism that involves IB␣ phosphorylation to trigger skeletal muscle cell differentiation.
IGF-II Induces PI3K-dependent IKK␣ and IKK␤ Activation during Myoblast Differentiation-The data presented above indicating that IB␣ phosphorylation was required by IGF-II to induce myoblast differentiation led us to analyze the activity and expression of IKK␣ and IKK␤ during this process. L6E9 myoblasts exhibited a peak of IKK␣ activity after 24 h in IGF-II-containing differentiation medium ( Fig. 2A, upper panel). The kinetics of IKK␣ activation was consistent with that of IGF-II-dependent NF-B DNA-binding activation, which we have previously shown to be dependent on PI3K activity (10). To analyze whether IGF-II-dependent IKK␣ activation involves PI3K, L6E9 myoblast differentiation was induced with or without IGF-II in the absence or presence of the PI3K inhibitor LY294002. The PI3K inhibitor (20 M) blocked the ability of IKK␣ to phosphorylate GST-IB␣-(1-54) in response to IGF-II (Fig. 2B, upper panel). Neither IGF-II nor LY294002 altered IKK␣ protein expression (Fig. 2C, upper panel); however, the 24-h delay between the start of the IGF-II treatment and the activation of IKK␣ seemed to indicate that IKK␣ was not the direct target of the IGF-induced phosphorylation cascade. This appears to be the case, as IKK␣ activation by IGF-II was blocked in the presence of 5 g/ml cycloheximide (Fig. 2D, upper panel, CH), indicating that it depends on de novo protein synthesis.
Phosphorylation of the GST-IB␣-(1-54) substrate by IKK␤ was also induced 24 h after triggering differentiation with IGF-II ( Fig. 2A, lower panel). IKK␤ activation by IGF-II was blocked by the PI3K inhibitor LY294002 (Fig. 2B, lower panel). Neither IGF-II nor LY294002 treatment modified the level of IKK␤ expression in differentiating myoblasts (Fig. 2C, middle panel), indicating that changes in activity are caused by activation of the kinase rather than variations in IKK␤ protein content. As for IKK␣, the activation of IKK␤ by IGF-II was blocked by cycloheximide (Fig. 2D, lower panel).
IGF-II Requires IKK␣, but Not IKK␤, to Induce Skeletal Muscle Cell Differentiation-To determine whether the activation of IKKs by IGF-II is functionally linked to myogenic differentiation, we generated replication-deficient adenoviral vectors expressing FLAG-tagged wild-type (adv/FLAG-IKK␣ and adv/FLAG-IKK␤) or dominant-negative (adv/FLAG-IKK␣ (K44A) and adv/FLAG-IKK␤(K44A) forms of IKK␣ and IKK␤. Subconfluent L6E9 myoblasts were transduced with the different recombinant adenoviruses using adv/GFP as a control; and 36 h later, differentiation was induced by placing the cells in an IGF-II-containing medium.
Expression and kinase activity of the transduced proteins were evaluated 24 h after inducing differentiation with IGF-II. Cells transduced with adv/FLAG-IKK␣ exhibited GST-IB␣-(1-54) phosphorylation activity in immunoprecipitates with an anti-FLAG monoclonal antibody, whereas no specific IB␣ phosphorylation activity was detected in immunoprecipitates from cells transduced either with adv/GFP or adv/FLAG-IKK␣(K44A) (Fig. 3A, upper panel). Under these conditions, similar levels of wild-type and dominant-negative forms of IKK␣ were expressed, as verified by immunoblotting using anti-FLAG antibody (Fig. 3A, lower panel). After 4 days of differentiation in the presence of IGF-II, cells overexpressing FLAG-IKK␣(K44A) remained highly undifferentiated, as measured by caveolin-3 and myosin heavy chain expression, compared with cells transduced with either adv/GFP or adv/ FLAG-IKK␣. The expression of the non-muscle-specific protein ␤-actin was not altered by FLAG-IKK(K44A) overexpression (Fig. 3B).
To analyze the role of IKK␤ activity in myoblast differentiation, cells were transduced with adv/FLAG-IKK␤ or adv/ FLAG-IKK␤(K44A). Both proteins were expressed to similar levels as detected on anti-FLAG immunoblots (Fig. 3C, lower  panel), and adv/FLAG-IKK␤(K44A)-transduced cells exhibited no IB␣ kinase activity on anti-FLAG immunoprecipitates (upper panel). In contrast to IKK␣, IKK␤ does not seem to play an essential role in myoblast differentiation, as cells transduced with adv/FLAG-IKK␤(K44A) expressed skeletal muscle-specific proteins as efficiently as non-transduced cells or cells transduced with adv/GFP or adv/FLAG-IKK␤ (Fig. 3D). At the morphological level, overexpression of FLAG-IKK␤(K44A) did not alter the ability of myoblasts to fuse in response to IGF-II (78% of nuclei in myotubes from a total of 1488 nuclei randomly counted) compared with non-transduced control cells (78%, total of 1616), cells transduced with adv/GFP (82%, total of  ). B, total cell extracts from L6E9 myoblasts transduced with adv/IB␣(S32A/S36A) (adv/IB␣AA) or adv/GFP and differentiated with IGF-II (40 nM) for 24 h were incubated with a 32 P-labeled NF-B probe and analyzed by electrophoretic mobility shift assay. C, expression of IB␣, myosin heavy chain (MHC), and caveolin-3 (Cav-3) was analyzed by immunoblotting of total cell lysates from non-transduced myoblasts (nt) or myoblasts transduced with adv/ GFP or adv/IB␣(S32A/S36A) and differentiated for 4 days with IGF-II. ␤-Actin content was analyzed as a control of relative amounts of proteins in each sample. D, L6E9 myoblasts non-transduced or transduced with adv/GFP or adv/IB␣(S32A/S36A) were grown to confluence and then allowed to differentiate in the presence of IGF-II for 4 days. Morphological differentiation was assessed by myotube formation. Images shown are representative of 15-25 microscope fields taken at random. The scale is the same for all panels. All infections were performed in duplicate, and the results shown are representative of three independent experiments.

FIG. 2. Endogenous IKK activation during IGF-II-induced myoblast differentiation.
A, cell extracts (50 g) from L6E9 myoblasts differentiated with IGF-II (40 nM) were immunoprecipitated (IP) with anti-IKK␣ (upper panel) or anti-IKK␤ (lower panel) antibodies (10 g) preadsorbed on protein G-Sepharose beads, and kinase assay was performed as described under "Experimental Procedures" using GST-IB␣-(1-54) as a substrate. Nonspecific kinase activity was analyzed using a nonimmune antibody (data not shown). B, kinase assay was performed as described for A with cells differentiated for 24 h in the absence or presence of IGF-II (40 nM) with or without LY294002 (20 M). C, IKK expression was analyzed by immunoblotting (IB) of cell lysates obtained from myoblasts differentiated for 24 h in the absence or presence of IGF-II with or without LY294002. The relative amounts of proteins in each sample were checked by expression of the non-musclespecific protein ␤-actin. D, subconfluent L6E9 myoblasts were differentiated for 24 h in DMEM in the absence or presence of IGF-II (40 nM) with or without cycloheximide (CH; 5 g/ml). Kinase assay was performed as described above. The results shown are representative of three to four independent experiments. 1661), or cells transduced with adv/FLAG-IKK␤ (80%, total of 1011) ( Fig. 4; arrows show large accumulations of nuclei in myotubes). Conversely, when FLAG-IKK␣(K44A) was overexpressed, only thin, spindle-like myotubes were formed compared with the large myotubes observed in non-transduced cells, cells transduced with adv/GFP, or cells transduced with adv/FLAG-IKK␣ (Fig. 4). Indeed, after 4 days in IGF-II-containing differentiation medium, only 25% of the nuclei (total of 1367) in cells overexpressing FLAG-IKK␣(K44A) were in myotubes with Ͼ10 nuclei. Under the same culture conditions, 77% of the nuclei from cells overexpressing FLAG-IKK␣ (total of 2987) were in myotubes with Ͼ10 nuclei. Taken together, these results suggest that IKK␣ plays a relevant role in IGF-II-dependent morphological and biochemical differentiation of skeletal muscle cells, whereas IKK␤ is not essential to this process.
NIK Is Activated in Differentiating Myoblasts-NIK is a common mediator in the NF-B signaling cascades, and IKK␣ has been reported to be a better substrate than IKK␤ for phosphorylation by NIK (24). The endogenous kinase activity and protein expression of NIK in differentiating L6E9 myoblasts were studied. Induction of NIK autophosphorylation activity was detected after 24 h in IGF-II-containing differentiation medium (Fig. 5A). As observed for NF-B DNA-binding activation (10) and IKK␣ and IKK␤ activities (Fig. 2B), the activation of NIK was totally blocked by LY294002, indicating that IGF-II requires PI3K to activate NIK in differentiating myoblasts (Fig. 5A). No changes were detected in NIK protein expression in response to IGF-II or LY294002 (Fig. 5B). As for IKK␣ and IKK␤, NIK activation by IGF-II was blocked by cycloheximide (Fig. 5C), indicating that it requires de novo protein synthesis.
Interestingly, LY294002 inhibited NIK activation only when it was present together with IGF-II during the whole 24-h treatment. In contrast, when LY294002 was added during the last 6 h or the last 1 h of IGF-II incubation (at 18 or 23 h of differentiation, respectively), no inhibition of NIK activity was observed (Fig. 5D). These results suggest that PI3K is most probably involved in the de novo synthesis of the factor(s) required for NIK activation rather than being a direct upstream element of the NIK-activating cascade.
To test whether the activation of NIK by IGF-II was a key event during differentiation, we transduced L6E9 myoblasts with recombinant adenovirus expressing kinase-proficient FLAG-tagged wild-type NIK that exhibited a high degree of basal autophosphorylation activity on anti-FLAG or anti-NIK immunoprecipitates ( Fig. 6A; the lower band that was detected with anti-NIK antibodies in cells transduced with adv/FLAG-NIK is probably due to cross-activation of endogenous NIK by the overexpressed kinase). After 2 days in the absence of IGF-II (DMEM), myoblasts infected with control adv/GFP fused poorly (12% from a total of 1713 nuclei counted at random were found in multinucleated myotubes) (Fig. 6B). In contrast, transduction with adv/FLAG-NIK promoted myoblast fusion (48% of nuclei in myotubes from a total of 1789 nuclei counted at random) (Fig. 6B). Co-immunofluorescence assays showed that NIK-overexpressing cells highly expressed myogenin in their nuclei (Fig. 6C). The number of nuclei expressing myogenin was 6-fold higher in cells overexpressing NIK (adv/NIK) than in control cells (adv/LacZ) (15 randomly selected fields were analyzed from each of two independent experiments with each condition performed in triplicate). DISCUSSION We describe a pathway by which IGF-II modulates skeletal muscle cell differentiation through activation of the IKK com- plex. The activation of the NF-B cascade (NIK, IKK␣, and IKK␤ activation; IB degradation; and NF-B DNA-binding activation) was detected 24 h after placing subconfluent myoblasts in IGF-II-containing differentiation medium. These are early events during the differentiation program induced by IGF-II, but they differ greatly in their kinetics compared with the rapid activation of this cascade, which can be detected within minutes of exposure of cells to cytokines or other stimuli. These observations indicate that IKK might not be a direct target of the IGF-induced phosphorylation cascade, an assumption reinforced by the fact that IGF-mediated IKK and NIK activation requires de novo protein synthesis. The nature of the factor(s) induced by IGF-II to trigger NF-B cascade activation during myogenesis remains undefined. Among the possibilities to be considered is that IGF-II could induce the secretion of an autocrine factor by differentiating myoblasts. In this context, we did not detect expression of tumor necrosis factor-␣ (a classical activator of NF-B) by reverse transcription-polymerase chain reaction studies in myoblasts induced to differentiate by IGF-II (data not shown). Another possibility is that the newly synthesized protein is a differentiation-induced kinase that may be a target for IGF-II itself or for an alternative factor generated in response to IGF-II.
The involvement of NF-B in myogenic signaling has been described in rat, human, and chick embryonic myoblasts (10,28,29), although differentiation of C2C12 skeletal muscle cells in a 2% horse serum-containing medium seems to occur through a signaling cascade in which NF-B plays a negative regulatory role (30). We have previously established that PI3K, NF-B, and inducible nitric-oxide synthase are elements of a common myogenic cascade in which IGF-II induces, through a PI3K-dependent pathway, a decrease in IB␣ protein content that correlates with a decrease in the amount of IB␣ associated with p65 NF-B, NF-B DNA-binding activation, and NO production (10). PI3K is a key mediator of myogenesis (5)(6)(7)(8)(9), and the role of PI3K in NF-B cascade activation during myogenesis is reinforced by data presented here showing that activation of both NIK and IKK by IGF-II in differentiating myoblasts is blocked by inhibiting PI3K. PI3K is known to be directly involved in the activation of NF-B in processes like anti-apoptotic platelet-derived growth factor signaling (31) and tumor necrosis factor-mediated immune and inflammatory responses (32). However, in view of our results, PI3K does not seem to directly activate NIK phosphorylation (LY294002 inhibited NIK activation only when it was present together with IGF-II during the whole 24-h treatment, but not during the last 6 h or the last 1 h). These data suggest that PI3K is most probably involved in the de novo synthesis of protein(s) required for NIK activation rather than being an upstream element in the activation of NIK. In contrast, our results do not rule out a biphasic role for IGF-II: first, in initiating the de novo synthesis of a NIK-activating factor and, second, in promoting NIK activation by the newly synthesized protein that (in view of the results presented in Fig. 5D) could occur only through a PI3K-independent pathway.
Although the stimuli that activate IKK␤ and the substrates that mediate its biological activity are known, the stimuli and the relevant substrates for IKK␣ are less well characterized. IKK␣ and IKK␤ appear to exert different and non-interchangeable physiological roles. Gene targeting experiments revealed that although IKK␣ is not involved in the activation of NF-B by pro-inflammatory stimuli, it is involved in morphogenesis (20 -22). In this context, our results show that although IGF-II induced both IKK␣ and IKK␤ activities early during the differentiation program, the overexpression of a kinase-deficient mutant of IKK␤ did not alter the expression of muscle-specific proteins or the formation of multinucleated myotubes. The differentiation process was, however, blocked by a kinase-deficient mutant of IKK␣, suggesting that endogenous IKK␤ cannot substitute for IKK␣ in myogenic signaling. Interestingly, skeletal muscle poorly expresses IKK␤, whereas it is one of the tissues with the higher expression levels of IKK␣ (33).
Proliferation precedes differentiation in IGF-stimulated myogenesis, and the opposing early and late effects of IGF during myogenesis are reflected in the phosphorylation state of the cell cycle regulatory retinoblastoma protein in skeletal myoblasts (34,35). Before exiting from the cell cycle, IGFs induce a last round of proliferation in which NF-B is required to increase cyclin D1 expression and pRb hyperphosphorylation (33,35,37). Then, a decrease in NF-B activity followed by a decrease in cyclin D1 levels seems to be required to allow the exit from the cell cycle that precedes differentiation (3, 36). A preadsorbed on protein G-Sepharose beads. Autophosphorylation assay was performed as described under "Experimental Procedures," and nonspecific (ns) NIK activity was tested using a nonimmune antibody. B, NIK expression was analyzed by Western blotting of cell lysates obtained at 24 h of differentiation, and the relative amounts of proteins in each sample were checked by expression of ␤-actin. C, subconfluent L6E9 myoblasts were differentiated for 24 h in DMEM in the absence or presence of IGF-II (40 nM) with or without cycloheximide (CH; 5 g/ml). Kinase assay was performed as described above. D, subconfluent L6E9 myoblasts were differentiated for 24 h with IGF-II (40 nM) with or without LY294002 (20 mM), added either for the whole 24 h or during the last 6 h or the last 1 h of IGF-II treatment. Kinase assay was performed at 24 h as described above. The data shown are representative of three independent experiments. causative relation between NF-B down-regulation and myogenesis was initially proposed, without considering that the NF-B activity detected after 24 h in differentiation medium (even if it was lower than that observed during proliferation) could exert a myogenic role (36). Indeed, this appears to be the case, as, consistent with data reported here, NF-B activity was shown to be required by human, mouse, and chicken myoblasts to fuse (10,31,32). We show here that IGF-II-dependent differentiation triggers a delayed induction of the NF-B-activating cascade, which requires PI3K activity and de novo synthesis of still undefined factors. Our data suggest that the activation of NIK and IKK␣ and the subsequent phosphoryla-tion of IB␣ at Ser-32 and Ser-36 are key events in skeletal muscle differentiation induced by IGF-II.
FIG. 6. NIK overexpression induces myoblast differentiation in the absence of IGF-II. A, 293 cells were transduced with adv/ FLAG-NIK. Non-transduced (nt) cells were used as controls. Cell extracts obtained 24 h after infection were immunoprecipitated (IP) with antibodies against NIK or FLAG preadsorbed on protein G-Sepharose beads. Autophosphorylation assay was performed as described under "Experimental Procedures." Nonspecific (ns) NIK autophosphorylation activity was tested using nonimmune antibodies. B, subconfluent L6E9 myoblasts were transduced with adv/FLAG-NIK or control adv/GFP. 36 h after infection, cells were allowed to differentiate for 2 days in the absence of IGF-II (DMEM). Morphological differentiation was assessed by myotube formation, and cells were photographed after nuclear staining. Images shown are representative of 25 microscope fields taken at random from each of three independent experiments. The scale is the same for both panels. C, shown is myogenin expression in L6E9 cells transduced with adv/FLAG-NIK or control adv/LacZ. Cells were grown to subconfluence on glass coverslips and maintained for 2 days in serum-free medium in the absence of exogenous IGF-II (DMEM). For immunofluorescence detection, cells were fixed and simultaneously probed for myogenin and NIK as described under "Experimental Procedures." Cells showing myogenin nuclear staining were counted and averaged from a minimum of 15 randomly selected fields. All infections were done in triplicate, and the results shown are representative of two independent experiments.