Histone deacetylase 6 is a FoxO transcription factor-dependent effector in skeletal muscle atrophy

Abstract

Skeletal muscle atrophy is a severe condition of muscle mass loss. Muscle atrophy is caused by a down-regulation of protein synthesis and by an increase of protein breakdown due to the ubiquitin-proteasome system and autophagy activation. Up-regulation of specific genes, such as the muscle-specific E3 ubiquitin ligase MAFbx, by FoxO transcription factors is essential to initiate muscle protein ubiquitination and degradation during atrophy. HDAC6 is a particular HDAC, which is functionally related to the ubiquitin proteasome system via its ubiquitin binding domain. We show that HDAC6 is up-regulated during muscle atrophy. HDAC6 activation is dependent on the transcription factor FoxO3a, and the inactivation of HDAC6 in mice protects against muscle wasting. HDAC6 is able to interact with MAFbx, a key ubiquitin ligase involved in muscle atrophy. Our findings demonstrate the implication of HDAC6 in skeletal muscle wasting and identify HDAC6 as a new downstream target of FoxO3a in stress response. This work provides new insights in skeletal muscle atrophy development and opens interesting perspectives on HDAC6 as a valuable marker of muscle atrophy and a potential target for pharmacological treatments.

Keywords: FOXO; Histone Deacetylase 6 (HDAC6); MAFbx; Muscle Atrophy; Proteasome; Skeletal Muscle; Ubiquitin.

FIGURE 1.

A, up-regulation of MAFbx (gray bars) and HDAC6 (black bars) in adult mice TA muscle induced by 3, 7, 10, 14 days of denervation. Quantitative PCR (QPCR) analysis was performed on 3 mice for each time point. Results are represented as fold induction compared with innervated. Error bars represent S.E. B, Western blot of extracts from adult mice TA muscle after 3, 7, 10, 14 days of denervation. Upper panel: HDAC6 blot. Lower panel: Coomassie staining as a total protein loading control. The experiment was repeated three times on 3 independent sets of mice. One representative experiment is shown. C, quantification of the fold increases of HDAC6 protein compared with the innervated control. Quantifications of 3 different Western blots are shown. Data are mean of three independent experiments. Error bars represent S.E. D, relative expression levels of HDAC6 were measured by QPCR in muscles biopsies of patients with muscle atrophy (patient 1 = p1, patient 2 = p2, patient 3 = p3, patient 4 = p4) or control muscles (c1). Results are normalized to the levels of HPRT expression. E–H, hematoxylin eosin staining of deltoid muscle cryosections performed on patients with muscle atrophy (p1, p2, p3, p4) and (I) control muscles (c1). Magnification ×200.

FIGURE 2.

A and B, muscles were electroporated either 2 days before (A) or 7 days (B) after denervation. 14 days later, cryosections were preformed and the cross sectional area of electroporated fibers was measured. Frequency histograms show the distribution of cross-sectional areas of muscle fibers of TA either innervated or 15 days denervated and electroporated with shHDAC6-GFP or control shRNA. As a control an shRNA against the luciferase was used. The median cross sectional area value is indicated for each condition. n = 5 mice for each condition. The asterisks indicate statistically significant changes at p < 0.05 (1 asterisk), calculated by Mann-Whitney's U test. den: 14 days denervated muscle; inn: innervated muscle. C, histogram shows the median values of the cross-sectional areas of TA fibers, electroporated with two different HDAC6 shRNA expressing vector (V1 and V2). D, Western blot showing the efficiency of V1 and V2 HDAC6 shRNA to reduce HDAC6 levels in C2C12. The experiment was repeated 3 times. One representative experiment is shown. E, quantification of panel D by Image J software. HDAC6 levels were normalized to tubulin levels.

FIGURE 3.

A, up-regulation of HDAC6 in FoxO triple KO and wild type mice TA muscles after 5 days of denervation. QPCR analysis was performed on 3 mice for each condition. Results are presented as fold induction compared with innervated. Error bars represent S.E. The asterisk indicates statistically significant changes at p < 0.01 (p = 0.004). B, consensus sequence for FoxO3 (GTAAAGA) at position +919 in the HDAC6 gene region. C, muscle extracts from denervated TA muscles were tested for binding to double stranded ³²P-labeled oligonucleotides containing a FoxO3 binding site by EMSA. Arrows indicate the FoxO3-DNA complex and the free DNA. Lane 1: free DNA alone; all other lanes contain the protein extract and, as indicated at the top of the figure, no oligonucleotide (positive control; lane 2), a competitor oligonucleotide (FoxO wt, FoxO mutant or NFkB; lanes 3–5) or an antibody (IgG or anti-FoxO3; lanes 6–7). D and E, localization of FoxO3 protein at the HDAC6 gene. Chromatin immunoprecipitation (ChIP) analysis were performed in wild type mice TAs muscles, either innervated, or 3 or 5 days denervated. Two indicated regions (A and B) were tested for QPCR amplification. QPCR values were calculated by extrapolation from a standard curve of Inputs DNA dilutions. Enrichment values were normalized to IgG signals and shown as the fold difference relative to region B. Each ChIP-QPCR histogram indicates the mean of triplicate results, error bars indicate S.E. A: FoxO binding site region at +919, B: negative region at +15074.

FIGURE 4.

A, indicated tagged proteins (Myc-HDAC6 and Flag-MAFbx) were expressed in C2C12 cells (input panel). Flag-MAFbx was then immunoprecipitated with an anti-Flag antibody and detected by Western blot with the same antibody (lower panels) and the co-immunoprecipitated Myc-HDAC6 was detected by Western blot using an anti-Myc antibody (upper panel). B, indicated tagged proteins expressing vectors (GST-HDAC6 and Flag-MAFbx) were electroporated in TA muscles (input panel). GST-HDAC6 was then immunoprecipitated with GST antibody conjugated Sepharose beads and detected with an anti-HDAC6 antibody (upper panels). The co-immunoprecipitated Flag-MAFbx was detected by Western blot using an anti-Flag (lower panels). C, indicated HA-tagged fragments of HDAC6 and Flag-MAFbx were expressed in C2C12 cells (input panel). Flag-MAFbx was then immunoprecipitated with an anti-Flag antibody and detected by Western blot with the same antibody. The coimmunoprecipitated HA-HDAC6 fragments were detected by Western blot using an anti-HA antibody (IP panel). Arrow asterisk, co-immunoprecipitated DD1. HDAC6 fragments are: N is the N-terminal non-catalytic domain, DD1 and DD2 respectively correspond to the catalytic domains 1 and 2 of HDAC6, and C corresponds to the ubiquitin-binding domain of HDAC6. D, indicated HA-tagged mutants of HDAC6 and Flag-MAFbx were expressed in C2C12 cells (input panel). Flag-MAFbx was immunoprecipitated with an anti-Flag antibody and detected by Western blot with the same antibody. Coimmunoprecipitated HA-HDAC6 mutants were detected by Western blot using an anti-HA antibody (IP panel). Arrow asterisk, co-immunoprecipitated N+DD1 and HDAC6. Arrow, absence of the co-immunoprecipitated ΔDD1. HDAC6 mutants are: HDAC6 is full-length HDAC6, N+DD1 is the N-terminal plus DD1, ΔDD1 is a full-length HDAC6 lacking the DD1.

FIGURE 5.

A, frequency histograms showing the distribution of cross-sectional areas of muscle fibers of TA muscles either innervated or 14 days denervated and electroporated with the HDAC6 enzymatic dead mutant or control plasmids. 14 days later, cryosections were preformed, and the cross sectional area of electroporated fibers was measured. The median cross sectional area value is indicated for each condition. n = 5 mice for each condition. The asterisks indicate statistically significant changes at p < 0.05 (1 asterisk), calculated by Mann-Whitney's U test. den: 14 days denervated muscle; inn: innervated muscle. B, C2C12 were cotransfected with and Flag-tagged MAFbx and either pcDNA expression vector, HA-tagged full-length HDAC6 or HDAC6 enzymatic dead (ED). Interactions were evaluated by immunoprecipitation with an anti-Flag antibody, followed by immunoblotting with an anti-Flag or anti-HA antibody. Total lysates used in immunoprecipitation are shown as input. C, HDAC6-shRNA is efficient in cultured cells. C2C12 myoblasts were transfected with either the HDAC6-shRNA or the control vector. 24 to 48 h after transfection, α-tubulin and HDAC6 were detected on protein extracts by Western blot. D, C2C12 cells were transfected either with a MAFbx expression vector, or with a MAFbx expression vector, and a HDAC6 shRNA vector, or with a MAFbx expression vector, a HDAC6 shRNA vector and a CMV-HDAC6 expression vector, to restore normal HDAC6 levels. 36 h after transfection, MyoD levels were evaluated by Western blot. E, C2C12 cells were cotransfected with vectors expressing HA-MyoD, MAFbx, and the HDAC6-shRNA as indicated. The cells were incubated with cycloheximide (CHX) 24 h after transfection, and total proteins harvested at the indicated times of treatment for analysis by Western blotting with anti-HA or anti α-tubulin.