IKK/NF-kappaB regulates skeletal myogenesis via a signaling switch to inhibit differentiation and promote mitochondrial biogenesis

Abstract

Nuclear factor kappaB (NF-kappaB) is involved in multiple skeletal muscle disorders, but how it functions in differentiation remains elusive given that both anti- and promyogenic activities have been described. In this study, we resolve this by showing that myogenesis is controlled by opposing NF-kappaB signaling pathways. We find that myogenesis is enhanced in MyoD-expressing fibroblasts deficient in classical pathway components RelA/p65, inhibitor of kappaB kinase beta (IKKbeta), or IKKgamma. Similar increases occur in myoblasts lacking RelA/p65 or IKKbeta, and muscles from RelA/p65 or IKKbeta mutant mice also contain higher fiber numbers. Moreover, we show that during differentiation, classical NF-kappaB signaling decreases, whereas the induction of alternative members IKKalpha, RelB, and p52 occurs late in myogenesis. Myotube formation does not require alternative signaling, but it is important for myotube maintenance in response to metabolic stress. Furthermore, overexpression or knockdown of IKKalpha regulates mitochondrial content and function, suggesting that alternative signaling stimulates mitochondrial biogenesis. Together, these data reveal a unique IKK/NF-kappaB signaling switch that functions to both inhibit differentiation and promote myotube homeostasis.

Figure 1.

Loss of p65 enhances myogenic activity in MEFs. (A) p65 ^+/+ and p65 ^−/− MEFs were cotransfected with cytomegalovirus-MyoD and either of the following reporter constructs: TnI-luc, AchR-luc, or 4RTK-luc. The next day, cells were switched to differentiation medium (DM), and, after 48 h, lysates were prepared and assayed for luciferase activity. (B) p65 ^+/+, p65 ^+/−, p65 ^−/− and pRb ^+/+, and pRb ^−/− primary MEFs were transfected with MyoD and TnI-luc. Cells were differentiated as in A, and luciferase assays were performed. (C) p65 ^−/−, cRel ^−/−, and p50 ^−/− MEFs were transfected with MyoD and TnI-luc, differentiated, and monitored for luciferase activity. (D) Myogenic assays similar to those described in A–C were performed in IκBα ^+/− and IκBα ^−/− cells. (inset) Electrophoretic mobility shift assay analysis of IκBα ^+/− and IκBα ^−/− MEFs. Error bars represent SEM.

Figure 2.

Loss of p65 accelerates the myogenic program in MEFs. (A) p65 ^+/+ and p65 ^−/− MEFs were infected with MSCV-MyoD and, after puromycin selection, were sorted for GFP to ensure equal MyoD levels. Cells were then probed for p65 and MyoD (45 kD) protein. α-Tubulin (55 kD) was used as a loading control. (B) p65 ^+/+ and p65 ^−/− MEFs stably expressing MyoD were differentiated, and lysates were then probed for the indicated myogenic differentiation markers. (C) Cells were differentiated as in B, and MyHC immunofluorescence was performed. (D) p65 ^−/− MEFs were transfected with TnI-luc and either vector plasmid, wild-type p65 (1–551), or p65 TA mutants (1–521 and 1–313) along with MyoD. Cells were then differentiated, and lysates were prepared for luciferase assays. RLU, relative light units. (E) p65 ^−/− MEFs were reconstituted with either vector and full-length or truncated p65 along with MSCV-MyoD. After selection, whole cell lysates were prepared and probed for p65, MyoD, and α-tubulin. (F) Cells were infected as in E, differentiated for 72 h, fixed, and stained for MyHC. (G) p65 ^−/− MEFs were transfected with MyoD, TnI-luc, and either vector control, wild-type p65 (WT), or p65 constructs containing S/A mutation at positions 276, 529, and 536. MEFs were differentiated and harvested after 48 h for luciferase assays. (H) Relative luciferase activities from p65 ^−/− MEFs transfected with MyoD, TnI-luc, and either vector control, wild-type p65, or p65 (1–313) containing the S276A mutation. Error bars represent SEM. Bars: (C) 200 μm; (F) 80 μm.

Figure 3.

Loss of p65 enhances the differentiation of primary myoblasts. (A) Primary myoblasts were prepared from 2–4-d-old TNFα ^−/−;p65 ^+/+, TNFα ^−/−;p65 ^+/−, and TNFα ^−/−;p65 ^−/− neonates, and genotypes were verified by Western blots for p65. (B) p65 and p50 primary myoblasts were transfected with TnI-luc or MyHC-luc plasmids, differentiated for 48 h, and subsequently harvested for luciferase assays. RLU, relative light units. (C) TNFα ^−/−;p65 ^+/+ and TNFα ^−/−;p65 ^−/− myoblasts were differentiated for 0 h (GM) or 48 h (DM) and subsequently stained for MyHC. (D) Quantification of myogenesis was performed by scoring MyHC-positive cells from a minimum of 25 fields normalized to total cell number as determined by Hoechst staining. (E) Myoblasts were differentiated for 0 (GM) and 48 h (DM), and lysates were probed for MyHC and Tn. The asterisk indicates Tn expression under GM conditions. (F) Primary or C2C12 myoblasts were transfected with siControl (siCont) or siRNA against p65 (sip65) along with Tn-luc reporter. Cells were switched to DM, and luciferase assays were performed. (G) C2C12 myoblasts were transfected with siControl or siRNA against p65 and switched to DM for 48 h, after which lysates were prepared and Western blots were performed. Error bars represent SEM. Bar, 80 μm.

Figure 4.

Myogenesis is enhanced in p65-deficient mice. (A) Hematoxylin- and eosin-stained cryosections of tibialis anterior (TA), gastrocnemius (Gastroc), and quadriceps (Quad) muscles from TNFα ^−/−;p65 ^+/+ and TNFα ^−/−;p65 ^−/− mice or p50 ^+/+ and p50 ^−/− gastrocnemius. (B) Fiber diameters were measured from gastrocnemius muscle sections from a total of 1,500 fibers (n = 5 mice per group). (C) Fiber numbers were determined in whole cross sections from tibialis anterior muscles from TNFα ^−/−;p65 ^+/+ and TNFα ^−/−;p65 ^−/− mice (n = 3). (D) Fiber numbers were recorded from premeasured randomly selected areas (minimum of 25 per animal) throughout the tibialis anterior, gastrocnemius, and quadriceps muscles (n = 5 mice per genotype). Error bars represent SEM. Bar, 50 μm.

Figure 5.

p65 regulation of myogenesis occurs through multiple mechanisms. (A) C2C12 myoblasts were transfected with control and p65 siRNA, and lysates were harvested for Western analysis. (B) MyoD ^−/− myoblasts were transfected with TnI-luc along with either an empty vector (CMV-Vect) or a p65 expression plasmid (CMV-p65) or were transfected with vector and subsequently treated with 5 ng/ml TNFα. Cells were differentiated for 2 d, and luciferase assays were performed. RLU, relative light units. (inset) Western blot for MyoD wild-type and null myoblasts. (C) MyoD ^−/− myoblasts were infected with pBabe-Puro or pBabe-p65 retroviruses. After selection and differentiation for 3 and 5 d, protein lysates were prepared for Western analysis. (D) MyoD ^−/− myoblasts stably expressing pBabe-Puro and pBabe-p65 were differentiated, fixed, and stained for MyHC (red) and nuclei (Hoechst; blue). (E) p65 ^+/+ and p65 ^−/− MEFs were transfected with TnI-luc and either MyoD or myogenin plasmids. As a control, transfections were also performed with a p53 expression plasmid and responsive reporter (pGL13-luc). Cells were subsequently differentiated, and luciferase assays were performed. Error bars represent SEM. Bar, 80 μm.

Figure 6.

IKK signaling is temporally regulated and functionally distinct in myogenesis. (A) IKKβ (f/f) MEFs or myoblasts prepared from E13.5 embryos or 3-d-old pups, respectively, were infected with pBabe-Puro or pBabe-Cre retrovirus. After selection, cells were transfected with MyoD and TnI-luc, and luciferase assays were performed after 2 d in DM. (B) Hematoxylin- and eosin-stained cryosections from tibialis anterior muscles of 4–6-wk-old IKKβ f/f and IKKβ f/f;muscle creatine kinase–Cre mice. (C) Fiber numbers were determined from premeasured randomly selected areas (minimum of 25 per animal) throughout the tibialis anterior muscle (n = 3 mice per genotype). *, P = 0.005. (D) IKK wild-type and null MEFs were transiently transfected with MyoD and TnI-luc, and, after 2 d in DM, lysates were prepared for luciferase assays. (E) C2C12 myoblasts were differentiated, and, at the indicated times, cells were harvested, and lysates were prepared for IKK kinase assays using wild-type or serine to alanine mutant IκBα proteins as substrates (KA, kinase assay; WB, Western blot). (F) C2C12 cells were differentiated, and, at the indicated time points, extracts were prepared to probe for phosphorylated IκBα, total IκBα, phosphorylated p65, and total p65. Parallel samples were prepared for nuclear extraction, and Western blots were performed for nuclear p65. Parallel differentiated C2C12 cells were immunoprecipitated with a p65 antibody and processed for chromatin immunoprecipitation (ChIP). Fragments from the IκBα promoter were amplified by PCR before (input) or after immunoprecipitation. (G) Lysates from differentiating C2C12 cells were prepared and used to probe for p100–p52 and α-tubulin. (H) MEFs wild type or null for p52 and RelB were transfected with MyoD and TnI-luc. Cells were differentiated for 48 h and prepared for luciferase assays. Error bars represent SEM. Bar, 0.5 μm.

Figure 7.

IKKα regulates myotube maintenance. (A) C2C12 cells were transfected with vector or HA-IKKα expression plasmids. After selection, cells were differentiated and harvested for Western analysis probing for HA and myogenic markers. (B) Myoblasts were transfected with siControl or siIKKα oligonucleotides and differentiated, and Western blotting was performed as in A. (C) 3-d differentiated myotubes stably expressing vector or IKKα were subjected to varying stress conditions, including no media replenishment for 6 d (6 d in DM) or low glucose (1 g/L glucose in DM for 48 h). Cells were then fixed and photographed by phase contrast at 20× magnification. (D) Differentiated myotubes stably expressing wild-type (WT) or a kinase-dead (KD) version of IKKα were switched to low glucose for 24 h and photographed by phase contrast. (E) Myotubes expressing siControl, siIKKα, or siIKKβ were differentiated for 3 d and switched to low glucose for 20 h before fixation. Parallel samples were harvested for Western blots to confirm knockdown efficiency. (F) C2C12 cells expressing vector or IKKα were differentiated for 6 d. Lysates were subsequently prepared for Western blots probing for IKKα and myogenic markers. Bars, 200 μm.

Figure 8.

IKKα regulates mitochondrial biogenesis. (A) DNA was prepared from GM or 3-d DM C2C12 cells, diluted, and used to amplify a 648-bp fragment from MTCO1. Separate PCR for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to normalize for loading. (B) C2C12 cells stably expressing vector (Vect) or IKKα were differentiated, and DNA samples were prepared for the determination of mitochondrial number as in A. (C) Vector and IKKα-overexpressing cells were differentiated for 3 d and stained for mitochondria with MitoTracker green. Staining was viewed by fluorescence at 20× magnification. (D) Mitochondrial and cytoplasmic extracts were prepared from HA-IKKα or IKKα-depleted myotubes, and lysates were probed for cytochrome c. (E) IKKα-expressing or -depleted cells were differentiated for 3 d and lysed, and ATP production was measured by luminescence. An ATP standard curve was generated in parallel to convert luminescence readings into [ATP] (*, P = 0.02; **, P = 0.001). (F) Vector and IKKα cells were differentiated, lysed, and prepared for a citrate synthase assay. All experiments were initiated with equal protein and were performed during the linear phase of the reaction to ensure adequate substrate amounts (*, P = 0.01). A parallel set of myotubes expressing vector or IKKα was cultured and switched to DM, and dehydrogenase activity was measured by the conversion of MTS tetrazolium into aqueous formazan. Readings were taken at 15 min before saturation (**, P = 0.001). (G) C2C12 cells were transfected with vector, p52–RelB, or p50–p65 and differentiated for 3 d. DNA was prepared for the determination of mitochondrial numbers as in A or was processed for the determination of ATP production as in E (*, P = 0.001; **, P = 0.03). Error bars represent SEM. Bar, 40 μm.

Figure 9.

IKKα controls mitochondrial structure. (A) Ultrathin sections from vector or IKKα-expressing myotubes were analyzed by EM at 18,500× direct magnification. (B) Myotubes expressing control or IKKα siRNA were sectioned and visualized by EM as in A. (C) Microarray analysis was performed on vector and IKKα-expressing myotubes using the murine MG430.20 Affymetrix chip. Genes up-regulated in IKKα myotubes as compared with vector were analyzed by L2L analysis (http://depts.washington.edu/l2l/) for statistically significant enriched biological processes. Bars, 500 nm.

Figure 10.

A model for IKK/NF-κB signaling and function in skeletal myogenesis. The model depicts different phases of myogenesis from proliferating myoblasts to differentiated myotubes. In proliferating myoblasts, classical NF-κB signaling mediated by IKKβ and IKKγ leads to the activation of p65 that binds DNA and regulates gene expression to inhibit myogenesis. During differentiation, classical NF-κB is down-regulated, whereas the alternative signaling becomes activated. Activation of alternative signaling occurs late in the myogenic program to regulate mitochondrial biogenesis and myotube maintenance.