Histone deacetylase 6 inhibition promotes microtubule acetylation and facilitates autophagosome-lysosome fusion in dystrophin-deficient mdx mice

Abstract

Aim:

Duchenne Muscular Dystrophy is a progressive muscle wasting disease caused by mutations in the dystrophin gene. Despite progress in dystrophin-targeted gene therapies, it is still a fatal disease requiring novel therapeutics that can be used synergistically or alternatively to emerging gene therapy. Defective autophagy and disorganized microtubule network contribute to dystrophic pathogenesis, yet the mechanisms by which microtubule alterations regulate autophagy remain elusive. The present study was designed to uncover possible mechanisms underpinning the role of microtubules in regulating autophagy in dystrophic mice.

Methods:

Mdx mice were also supplemented with Tubastatin A, a pharmacological inhibitor of histone deacetylase 6, and pathophysiology was assessed. Mdx mice with a genetic deletion of the Nox-2 scaffolding subunit p47^phox were used to assess redox dependence to tubulin acetylation.

Results:

Our data show decreased acetylation of α-tubulin with enhanced histone deacetylase 6 expression. Tubastatin A increases tubulin acetylation and Q-SNARE complex formation but does not alter microtubule organization or density, indicating improved autophagosome-lysosome fusion. Tubastatin A increases acetylation of peroxiredoxin and protects it from hyper-oxidation, hence modulating intracellular redox status in mdx mice. Tubastatin A reduces muscle damage and enhances force production. Genetic down regulation of Nox2 activity in the mdx mice promotes autophagosome maturation but not autolysosome formation.

Conclusion:

Our data highlight that autophagy is differentially regulated by redox and acetylation in mdx mice. By improving autophagy through promoting tubulin acetylation, Tubastatin A decreases the dystrophic phenotype and improves muscle function, suggesting a great potential for clinical translation and treating dystrophic patients.

Introduction

Duchenne Muscular Dystrophy (DMD) is the X-linked recessive genetic disorder caused by mutations in DMD gene which encodes for dystrophin (Dp427m), a key component of the dystrophin-glycoprotein complex (DGC). It affects approximately 1:4000 to 1:5000 live male births worldwide¹ resulting in progressive muscle wasting and degeneration, leading to death due to cardiac dysfunction and respiratory failure.² Over the past two decades, several therapeutic approaches have been evaluated to combat the pathogenesis of the disease, but DMD is still incurable. Novel genetic approaches hold promising therapy;³ however, many challenges still exist due to the variability in exon skipping efficiency among patients, the non-homogenous restoration of dystrophin between muscle types (including absence of effect of current exon skipping agents on cardiac muscle), and minimal alterations in immune cell infiltration. To combat these limitations, the treatment plan for DMD will likely entail a combination of genetic and pharmacological interventions.

Dystrophin is a large cytoskeletal protein located at the sarcolemma that mechanically links the internal cytoskeleton to the extracellular matrix and is critical for muscle-membrane stability during contraction.⁴ Lack of dystrophin results in disassembly of the DGC and increases the sarcolemma susceptibility to contraction-induced injury.⁵ This leads to a series of pathological events including, increased ROS signaling, aberrant Ca²⁺ release, inflammation, impaired autophagy, fibrosis, apoptosis, and decreased force production.^{5, 6, 7}

Macroautophagy, hereafter referred to as autophagy, is a highly conserved process involving a series of sequential events for bulk degradation of cytosolic components and organelles through delivery of autophagosomes to lysosomes, and thus, maintaining cellular homeostasis.⁸ Accumulating evidence shows that impaired autophagy contributes to muscle weakness and cell death in both mdx mice (mouse model of DMD) and DMD patients.^{9, 10} Recent work from our lab has shown that Nox2/Src kinase impairs autophagy by regulating the PI3K/Akt/mTOR pathway in mdx mice.⁷ To better comprehend how defective autophagy leads to DMD pathophysiology, and to develop therapeutic strategies, we need to elucidate the defects at different steps of the autophagic pathway.

Dystrophin is a microtubule-associated protein found to bind microtubules (MT).¹¹ The MT lattice becomes disorganized when dystrophin expression is ablated as in the mdx mouse.^{11, 12, 13} We have shown that the increased Nox2-ROS observed in mdx skeletal muscle regulates the MT network.¹⁴ MTs undergo post-translational modifications that regulate their biological functions. Some studies suggest detyrosination of α-tubulin increases muscle stiffness and decreases force production in mdx mice.^{15, 16} In non-muscle cells, acetylation of α-tubulin has been shown to regulate the formation of pre-autophagosomal structures, vesicular movements and autophagosome-lysosome fusion.^{17, 18, 19} Acetylated MTs recruit the motor protein kinesin-1 to transport autophagosomes in a cargo-specific manner along the MT tracks. Changes in MT acetylation leads to alterations in MT dynamics and organization, cell migration, and autophagy.^{18, 20} Despite the extensive investigations largely based on MT dynamics, stability and its altered network in dystrophic mice, no study has addressed the role of acetylated MTs specifically on autophagosome biogenesis and autophagosome-lysosome fusion in DMD. Therefore, the present study was designed to unravel the plausible mechanisms underpinning the role of MTs in regulating autophagy in dystrophic mice.

Histone deacetylases (HDACs) are a class of deacetylase enzymes involved in chromatin remodeling and gene expression.²¹Epigenetic drugs (e.g. Givinostat) targeting HDACs have been studied^{2, 22} and are now FDA approved for use in DMD patients. Recently, HDAC6 (Class IIb) inhibition has emerged as one potential selective pharmacological target in neurodegenerative diseases. HDAC6 catalyzes the deacetylation of non-histone proteins such as α-tubulin, leading to altered MT stability and organization.²³ In addition, HDAC6 can control redox regulation through acetylation of peroxiredoxin (Prx1 and PrxII).^{24, 25} Therefore, selective HDAC6 inhibition has the potential to reduce the toxicity related to the off-target effects of pan-HDAC inhibitors (Givinostat).²⁶ Recent investigations have suggested that the HDAC6 inhibitor, Tubastatin A (TubA) promotes MT acetylation, improving autophagic flux, redox balance, and functional recovery in neurodegenerative disorders,^{19, 27, 28, 29} cardiomyopathy,³⁰ idiopathic pulmonary, fibrosis³¹ myocardial ischemia/reperfusion injury,³² osteoarthritis,³³ and kidney injury.³⁴ In the current study, we show differential regulation of autophagy in mdx skeletal muscle; while autophagosome maturation is regulated by Nox2-ROS, autophagosome-lysosome fusion is regulated by MT acetylation. Furthermore, we discovered the therapeutic efficacy of HDAC6 inhibitor, TubA, in promoting MT acetylation and improving autophagosome-lysosome fusion in mdx mice. TubA treatment significantly restricted muscle damage and apoptosis and induced muscle functional recovery.

Results:

Impaired autophagosomal biogenesis/maturation in mdx mice

The maturation of double-membrane vesicle structures called autophagosomes occurs in a highly orchestrated manner to achieve successful delivery to the lysosomes for fusion.³⁵ One of the initial steps in autophagosome biogenesis is the recruitment and activation of the class III phosphatidylinositol 3-kinase complex (PI3K), consisting of Beclin, ATG14L, VPS34, and VPS15 to facilitate the phagophore nucleation.³⁶ Immunoblot analyses of TA muscle homogenate revealed a significant decrease in ATG14L and VPS34, whereas both Beclin and VPS15 were found to be increased in mdx muscles as compared to WT (Figure 1A, B). Immunoprecipitation of Beclin followed by western blot for ATG14L showed decreased ATG14L/Beclin complex formation in mdx muscle as compared to WT while Bcl-2/Beclin complex formation was not altered (Figure 1C). These data are consistent with inhibition of vesicle nucleation in mdx mice. Autophagy induction by heterodimerization of Beclin with ATG14L-VPS34 is regulated by activated JNK. JNK-interacting protein 1 (JIP-1) binds protein kinases (e.g. MAPKK/MKK7) to promote phosphorylation and activation of JNK.³⁷ We observed that the protein expressions levels of JIP-1 and p-JNK (Thr183/Ty185) are downregulated in mdx skeletal muscle, whereas MAPKK (MEK-1a/MEK-2) was significantly upregulated in mdx mice (Figure 1D, E). P-JNK phosphorylates Bcl-2 (S70), which disrupts the Bcl-2/Beclin complex, allowing Beclin to bind the PI3KclassIII complex for phagophore nucleation. We found lower levels of p-Bcl-2/Bcl-2 in mdx muscles as compared to WT, likely due to the decrease in p-JNK levels (Figure 1D, E). Together, these observations suggest that phagophore nucleation is disrupted due to inhibition of JIP-1/JNK activation and subsequent phosphorylation of Bcl-2 in mdx mice.

Figure 1. Impaired autophagosomal biogenesis/maturation in mdx mice.

(A-B) Immunoblot and densitometry analysis for autophagy regulatory proteins involved in vesicle nucleation (n=6 per group) and (ATG14L; WT, n = 6; mdx, n = 5). (C) Skeletal muscle tissue lysate prepared and subjected to immunoprecipitation using anti-Beclin, and the associated ATG14L, and Bcl-2 were determined using immunoblotting (n=3 per group). IgG served as a negative control. (D-E) JIP-1/JNK signaling was detected by immunoblot with antibodies as indicated (n = 6 per group) and (p-Bcl-2; WT, n=6; mdx,n=5). Densitometry analysis of immunoblots represented by box plots. (F-G) Immunoblot for proteins involved in autophagosome maturation and elongation (WT, n = 6; mdx, n = 5). Densitometry analysis of immunoblot represented by graph. GAPDH served as a loading control. GAPDH shown for VPS34 (a) is the same as shown for JIP-1 (d), single SDS-PAGE gel was loaded with protein samples for VPS34 & JIP-1). Representative bands with dividing lines were cropped from the same blot. *p < 0.05, †p < 0.01, ‡p < 0.001, statistical difference between two groups were determined using two-sample student-t-test and Welch’s correction test was performed for non-equal variance between two groups.

Vesicle nucleation is followed by elongation and expansion of the phagophore in the cytoplasm, which is regulated by the ATG5–12 complex. Our data revealed that WIPI-1 (WD Repeat Domain Phosphoinositide Interacting 1), an early marker of autophagosome formation which fosters the recruitment of downstream ATGs, was significantly downregulated in mdx muscle (Figure 1F, G) whereas ATG7, ATG4B, and ATG5–12 did not show any prominent change in mdx as compared to WT (Figure 1F, G). At the mRNA level, gene expression analysis of autophagy related markers Uvrag, Vps34, Atg14l, FIP200, Atg5, Atg12, and Gabarapl1 did not show any change in mdx skeletal muscle as compared to WT (Figure S1). Together, our data suggests that defects in the autophagosome maturation are due to inhibition of vesicle nucleation and elongation of the phagophore in mdx skeletal muscle.

Impairment in the SNARE-mediated autophagosomal-lysosomal fusion in mdx mice

To clear sequestered cytosolic components the autophagosome must fuse with the lysosome, forming the autolysosome. The process of autophagosome-lysosome fusion is regulated by the SNARE tertiary complex STX17-SNAP29-VAMP8. To achieve autophagosome-lysosome fusion, autophagosome-localized Q-SNARE STX17 must interact with SNAP29 and the lysosomal-localized R-SNARE VAMP8. ATG14L acts as a tethering factor to facilitate the fusion by directly interacting with the STX17-SNAP29 binary complex, and primes it for binding to VAMP8 on lysosomes.³⁸ We found that STX17 was significantly decreased in mdx as compared to WT, while VAMP8 showed a trend to increase but did not reach statistical significance, and SNAP29 showed no change (Figure 2A, B). CO-IP with anti-SNAP29 in TA muscle lysates revealed reduced interaction of SNAP29 with STX17 and VAMP8 in mdx muscle whereas SNAP29 retained binding with ATG14L (Figure 2C). CO-IP with an anti-VAMP8 showed less interaction of VAMP8 with STX17, SNAP29, and ATG14L (Figure S2A), confirming reduced autophagosome-lysosome fusion in mdx skeletal muscle.

Figure 2. Impairment in the SNARE complex-mediated autophagosomal-lysosomal fusion in mdx mice.

(A-B) Protein expressions of SNARE tertiary complex-STX17, VAMP8, and SNAP29 was determined by immunoblot (WT, n = 6; mdx, n = 5). Densitometry analysis of immunoblot represented by box plots. (C) Skeletal muscle lysate was prepared and subjected to immunoprecipitation using anti-SNAP29, and the associated SNAP29, STX17, ATG14L, and VAMP8 were determined using immunoblotting (n=3 per group). IgG served as a negative control. GAPDH served as a loading control. GAPDH shown for SNAP29 (A) is the same as shown for ATG4B, (Figure.1F, single SDS-PAGE gel was loaded with protein samples for ATG4B & SNAP29). †p < 0.01, statistical difference between two groups were determined using two-sample student-t-test and Welch’s correction test was performed for non-equal variance between two groups.

Acetylation of STX17 inhibits the interaction between STX17 and SNAP29, and formation of the Q-SNARE complex.³⁹ Therefore, we asked whether STX17 acetylation is impaired in mdx muscle, leading to reduced interaction between the Q-SNARE complexes. We observed deacetylation of STX17 (Figure S2B) with increased expression of HDAC2 (Figure S2C–D), the primary deacetylase of STX17³⁹ in mdx muscle compared to WT. These data suggest that deacetylation of STX17 is not sufficient to allow STX17-SNAP29-VAMP8 SNARE complex assembly in mdx muscle.

Genetic ablation of p47^phox promotes autophagosomal maturation without facilitating autolysosome formation in mdx mice

We have previously shown that genetic deletion of p47^phox (encoded by the Ncf1 gene) function in mdx mice protects against oxidative stress and improves autophagy (i.e. increased LC3II/I and decreased p62) compared with mdx.⁷ Having established defects at multiple stages in autophagy, including autophagosome maturation and autolysosome formation in mdx mice, we asked whether these steps are regulated by Nox2-ROS in mdx skeletal muscle. We found increased JIP-1, p-JNK, and p-Bcl-2 protein levels (Figure 3A, B) in muscle from p47 KO/mdx compared to muscle from mdx mice, as well as increased Beclin-ATG14L complex formation and a decrease in the Beclin-Bcl-2 complex in p47 KO/mdx compared to mdx (Figure 3C); all consistent with improved vesicle nucleation. We also found an increase in the autophagy elongation complex, ATG5–12 in p47 KO/mdx mice (Figure 3D, E). We then determined whether autophagosome-lysosome fusion is affected by genetic deletion of p47^phox in mdx mice. Interestingly, we did not find any improvement in the interaction between STX17-SNAP29-VAMP8 tertiary complexes in p47 KO/mdx mice as compared to mdx mice (Figure 3F). In addition, no changes were found in the lysosomal protein LAMP2 and lysosomal hydrolase, Cathepsin B in p47 KO/mdx mice as compared with mdx (Figure 3G, H). Overall, our data indicate that inhibition of Nox2-ROS promotes autophagosome maturation but does not enhance the fusion of autophagosomes with lysosomes in mdx mice.

Figure 3. Genetic ablation of p47^phox promotes autophagosomal maturation in mdx mice.

(A-B) Immunoblot for JIP-1/JNK signaling proteins which participates in vesicle nucleation. Densitometry analysis of western blots represented by box plots (C) Skeletal muscle lysate was prepared from muscles of WT, mdx and p47 KO/mdx mice and subjected to immunoprecipitation using anti-Beclin, and the associated ATG14L, and Bcl-2 were determined using immunoblotting. IgG served as a negative control. Immunoblot for WT and mdx group are the same as shown in Figure 1C. (D-E) Protein expressions of autophagosome elongation marker, ATG5–12 was determined by western blot. Densitometry analysis of western blot represented by box plots. (F) Skeletal muscle tissue lysate was prepared from muscles of WT, mdx and p47 KO/mdx mice and subjected to immunoprecipitation using anti-SNAP29, and the associated SNAP29, STX17,, and VAMP8 were determined using immunoblotting. Immunoblot for WT and mdx group are the same as shown in Figure 2C. (G-H) Immunoblots of lysosomal proteins, LAMP2 and Cathepsin-B. Densitometry analysis of western blot represented by box plots. GAPDH served as a loading control. (n=3 per group in all the figure panels). Representative bands with dividing lines were cropped from the same blot. *p < 0.05, †p < 0.01, ‡p < 0.001, statistical difference between groups were determined using ANOVA with Tukey’s post hoc test.

Altered MTs acetylation affects autophagy

Emerging evidence shows that reversible acetylation of α-tubulin can regulate MT function, thus, facilitating fusion of autophagosomes to lysosomes.^{40, 41} The acetylation status of microtubules is coordinated by deacetylases (HDAC6)²³ and acetyltransferases (MEC-17).⁴² Immunoblot analyses showed that the α-tubulin acetylation levels were significantly decreased in mdx muscles as compared to WT (Figure 4A, B). The deacetylase enzyme HDAC6 was increased in mdx muscle (Figure 4A, B) whereas the acetyltransferase (MEC17) did not exhibit any change in mdx muscle (Figure S3A, B). At the mRNA level, expression of Hdac6 and Mec17 genes did not show any change in mdx as compared to WT (Figure S3C, D). Expression of Sirtuin 2 (SIRT2), a class III HDAC that can deacetylate microtubules,⁴³ was not different in skeletal muscle from mdx mice compared to WT (Figure S3A, B). Autophagosome movement along MT tracks is dependent upon the microtubule motor proteins kinesin and dynein. Kinesin-1 (conventional kinesin or KIF5), is a tetramer of two kinesin heavy chains (KHC or KIF5B) and two kinesin light chains (KLC)⁴⁴. KLCs play dual roles, they direct cargo binding and regulate motor activity. We found that KIF5B was significantly increased in mdx muscles whereas KLC and dynein did not change in mdx as compared to WT (Figure 4C, D). In addition, we did not find any restoration of acetylated α-tubulin levels in p47 KO/mdx mice (Figure 4E, F). Overall, our data strongly suggests that microtubule acetylation plays an important role in autophagosome-lysosome fusion in mdx mice, which does not appear to be regulated by Nox2-ROS.

Figure 4. Altered tubulin-acetylation in mdx muscle.

(A-B) Protein expressions of α-tubulin, acetylated-α-tubulin, HDAC6 were determined by western blot (n=6 per group). Densitometry analysis of immunoblot represented by box plots. (C-D) Protein expressions of MTs motor proteins, kinesin-1 complex (KIF5B, KLC) and dynein were determined by western blotting. WT and mdx mice (n=6 per group). Densitometry analysis of immunoblot represented by box plots (E-F) Protein expressions of α-tubulin, acetylated-α-tubulin, HDAC6 were determined by western blot in mdx and p47 KO/mdx mice (n=3 per group). GAPDH as a loading control. GAPDH shown for Dynein (C) is the same as shown for p-JNK (Figure1D, single SDS-PAGE gel was loaded with protein samples for Dynein & p-JNK). *p < 0.05, ‡p < 0.001, statistical difference between two groups were determined using two-sample student-t-test.

HDAC6 inhibition improves MT acetylation without altering MT organization in mdx mice

Acetylated microtubules play an important role in vesicle trafficking and fusion.^{18, 20} To assess the relationship of HDAC6 activity with microtubule alterations and impaired autophagic flux, HDAC6 was inhibited with its specific pharmacological inhibitor, Tubastatin A (TubA). TubA was intraperitoneally injected for 2 weeks in 3 week old mdx mice, which is just before the onset of disease progression. The dose given to the mdx mice was 70mg/kg per day which is equivalent to 8.4mg/kg per day for a human child based on FDA approved mouse to human-equivalence dose calculation guide.⁴⁵. TubA restored α-tubulin acetylation in mdx mice without altering the protein expression levels of either de-tyrosinated α-tubulin or HDAC6 (Figure 5A, B) nor kinesin, JIP-1, or JNK (Figure S4A, B). Immunofluorescence staining further confirmed the enhanced expression and localization of acetylated-α-tubulin within myofibers of TubA treated mdx mice (Figure 5C, D). This dual confirmation provides robust evidence that HDAC6 inhibitor, TubA restores acetylation levels in mdx mice. Furthermore, immunofluorescence staining of KIF5B corroborated the western blot findings, showing no further increase in the expressions of protein within myofibers of TubA-treated mdx mice (Figure S4C, D). Interestingly, localization of KIF5B was found around the central nuclei within the degenerating/regenerating myofibers of both mdx and mdx-treated TubA group. Altogether, our data suggests that HDAC6 inhibition enhanced tubulin acetylation without affecting the motor proteins to facilitate the autophagosome-lysosome transport. Next, we assessed whether other HDAC6 substrates, including cortactin, β-Catenin, and HSP-90, undergo any alterations in mdx mice. We found no significant differences in the acetylation levels of cortactin, β-catenin or HSP90 in mdx muscle compared to WT, although acetylation levels of β -catenin trended to be reduced in mdx muscle (Figure S5A, B). Furthermore, mdx mice treated with TubA, did not show any discernible changes in the acetylation levels of these protein compared to mdx mice (Figure S5). We found increased MT disorganization and density in flexor digitorum brevis muscle from mdx mice compared to WT and HDAC6 inhibition with TubA did not restore MT organization or density compared to mdx mice (Figure 5E–F). We also assessed TubA dosing in mdx cohorts at 30mg/kg every other day for 4 weeks and did not observe any muscle functional recovery, although serum CPK levels were reduced. (Figure S5C, D).

Figure 5. HDAC6 inhibition promotes tubulin acetylation without altering MT organization in mdx mice.

(A-B) Protein expressions of α-tubulin, acetylated-α-tubulin, detyrosinated α-tubulin, and HDAC6 were determined by western blot in WT, mdx and TubA treated mdx mice (n=3 per group). Densitometry analysis of immunoblot represented by box plots. (C-D) Representative immunofluorescence micrograph of anti-acetylated-α-tubulin labeled with secondary antibody Alexa Fluor 488 (Green) and total α-tubulin labeled with secondary antibody Alexa Fluor 594 (Red), counterstained with DAPI (nuclear stain) from section of gastrocnemius (GAS) muscle from WT (n=3), mdx (n=3) and TubA treated mdx (n=3)(scale bar −50μm; Magnification-40X). Quantitative analysis represented by box plot. (E) Representative images of FDBs fibers isolated from WT, mdx and mdx+TubA and stained with α-tubulin antibody conjugated with Alexa Fluor 488 to label MT network (Scale Bar-10μm; Magnification-40X). (F) Graph represents MT grid-like organization and density which were analyzed from five fibers from four to five mice (N_animals=4–5, N_fibers=20–25) using TeDT software and ImageJ respectively. GAPDH served as a loading control. *p < 0.05, †p < 0.01, ‡p < 0.001, statistical difference between groups were determined using ANOVA with Tukey’s post hoc test. *p<0.05 vs mdx and mdx+TubA, # p<0.05 vs mdx+TubA for panel F.

HDAC6 inhibition promotes autophagosome-lysosome fusion in mdx muscle

We assessed whether the increased acetylation of α-tubulin improved autophagic flux in mdx muscle. The autophagosome substrate p62 (SQSTM1) was significantly decreased in TubA treated mdx muscle while the ratio of LC3II/I remained elevated in TubA treated mdx mice due to a sustained elevation of LC3II (Figure 6A, B). The decreased p62 is consistent with an increase in cargo recognition and efficient degradation within autolysosomes.

Figure 6. HDAC6 inhibition promotes autophagosome-lysosome fusion in mdx muscle.

(A-B) Immunoblot for autophagy related proteins- p62 and LC3 (n = 3 per group). Densitometry analysis of western blots represented by box plots. (C) Skeletal muscle lysate was prepared and subjected to immunoprecipitation using anti-SNAP29, and the associated, STX17, ATG14L, and VAMP8 were determined using immunoblotting (n=3 per group). (D-E) Representative immunofluorescence micrograph of anti-SNAP29 labeled with secondary antibody Alexa Fluor 594 (Red) and anti-VAMP8 labeled with secondary antibody Alexa Fluor 488 (Green), counterstained with DAPI (nuclear stain) from section of gastrocnemius (GAS) muscle from WT (n=3), mdx (n=3) and TubA treated mdx (n=5)(scale bar −50μm; Magnification-40X).Quantitative values for colocalization analysis was represented as the Pearson’s Correlation Coefficient of SNAP29-VAMP8. (F-G) Protein expressions of lysosomal proteins, LAMP2 (n=3 per group) and Cathepsin-B (n=6 per group) were determined by western blot in WT, mdx, and TubA treated mdx mice. Densitometry analysis of western blots represented by box plots. GAPDH served as a loading control. Representative bands with dividing lines were cropped from the same blot. *p < 0.05, †p < 0.01 statistical difference between groups were determined using ANOVA with Tukey’s post hoc test.

To further examine the mechanism behind the efficient autophagic clearance in TubA treated mdx, we evaluated the autophagosome-lysosome fusion by detecting the binding of SNARE complex proteins. TubA treatment enhanced the interaction of SNAP29 with STX17 and VAMP8 (Figure 6C). Immunostaining of SNAP29 and VAMP8 showed puncta within the muscle and increased colocalization in TubA treated mdx muscle as compared to non-treated, indicative of enhanced autophagosome-lysosome fusion by HDAC6 inhibition (Figure 6D, E). In addition, the lysosomal protein LAMP2 and cleaved (active) lysosomal hydrolase cathepsin-B are elevated in TubA treated mdx muscle (Figure 6F, G). Chloroquine (CQ), an autophagy inhibitor, increased accumulation of p62 aggregates in mdx and TubA treated mdx skeletal muscle (Figure S6A–C). Overall, our data suggest that the decreased acetylation of α-tubulin in mdx muscle inhibits autophagosome-lysosome fusion, which can be recovered upon HDAC6 inhibition.

HDAC6 inhibition recovers acetylation of PrxII and increases total PrxII in mdx mice

Prx I and Prx II are specific substrates of HDAC6, their acetylation status provides resistance to hyperoxidation. We have previously shown hyperoxidation and proteolytic degradation of PrxII in mdx skeletal muscle.⁴⁶ Here, we immunoprecipated PrxII and probed for acetyl lysine and found decreased acetylated PrxII levels in mdx mice, which was recovered back to WT in TubA treated mdx mice (Figure 7A). This was further confirmed by immunoprecipitating acetyl lysine followed by probing for PrxII (Figure S7B). Total PrxII levels were found to be significantly increased in TubA-mdx mice (Figure 7B–E) while oxidized PrxII (PrxSO_2/3) showed no difference as compared to mdx (Figure 7B, C). Our findings that TubA did not decrease PrxSO_2/3 is not surprising, as sulfonation of PrxII is an irreversible oxidative modification.⁴⁷ We did find that the ratio of oxidized PrxII to total PrxII is significantly reduced in TubA-treated mdx (Figure 7B, C), indicating that the increased acetylation of PrxII prevents its hyperoxidation. To our surprise, we found that stretch activated ROS production was not diminished in diaphragm muscle from TubA treated mdx compared to mdx (Figure 7F). Intriguingly, acetylated PrxII levels were restored back to WT levels in p47 KO/mdx mice (Figure S7A, C). Total PrxII was increased, PrxSO_2/3 decreased, and the ratio of oxidized to total PrxII decreased in p47 KO/mdx as compared to mdx. (Figure S7D, E). These data suggest that either reducing oxidative stress (p47 KO/mdx) or increasing acetylation of PrxII (TubA) prevents its hyperoxidation and degradation in mdx skeletal muscle.

Figure 7. HDAC6 inhibition prevents deacetylation and hyperoxidation of PrxII in mdx mice.

(A) Skeletal muscle lysate was prepared and subjected to immunoprecipitation using anti-PrxII, and the acetylation levels of PrxII was determined with the anti-acetyl lysine antibodies (B-C) Protein expressions of total PrxII and Prx-SO_2/3 were determined by western blot in WT, mdx, TubA treated mdx mice (n=3 per group). Densitometry analysis of western blots and the ratio of PrxSO_2/3 and total PrxII were represented by box plots. (D-E). Representative immunofluorescence micrograph of total PrxII labeled with secondary antibody Alexa Fluor 488 (Green), counterstained with DAPI (nuclear stain) from section of gastrocnemius (GAS) muscle from WT (n=3), mdx (n=3) and TubA treated mdx (n=3) (scale bar −50μm; Magnification-40X). Quantitative analysis represented by box plot. (F) Stretch-induced ROS measurements in diaphragm muscles of mdx (n=3) and TubA treated mdx (n=3) mice. GAPDH served as a loading control. GAPDH shown for Prx-SO_2/3 (B) is the same as shown for DT-Tubulin (Figure 5A, single SDS-PAGE gel was loaded with protein samples for HDAC6, DT-tubulin & Prx-SO_2/3). *p < 0.05, †p < 0.01, statistical difference between groups were determined using ANOVA with Tukey’s post hoc test. Statistical difference between two groups were determined using two-sample student-t-test and Welch’s co

TubA treatment alleviates apoptosis and immune cell infiltration in mdx muscle

Impaired autophagy is associated with aggregation of oxidized/misfolded proteins and other cellular constituents, eventually leading to cell death. Immunoblot expression of cleaved/active caspase-3 was found to be significantly reduced in mdx mice following treatment with TubA (Figure 8A, B). The percentage of TUNEL positive nuclei (green, marked by white arrows) were significantly elevated in mdx gastrocnemius (GAS) muscle, which were significantly decreased upon treatment with TubA (RI_TubA =92.14%, Figure 8C, D). Infiltration of macrophages and other immune cells is an important and pathogenic feature of dystrophin-deficient muscle, even at asymptomatic stage of disease progression. We have stained GAS muscle cross-sections with anti-CD68 antibody to identify CD68+ macrophages. We found a significant decrease in CD68⁺ immune cells (green, marked by white arrows) in TubA treated mdx skeletal muscle (RI _TubA=63.10%, Figure 8E, F), indicating decreased infiltration of macrophages in the endomysium of skeletal muscles.

Figure 8. TubA treatment alleviates apoptosis and immune cells infiltration in mdx muscle.

(A-B) Protein expressions of pro-caspase-3 and cleaved caspase-3 were determined by western blot from TA muscles extract of WT, mdx, and TubA treated mdx (n = 3 per group). Densitometry analysis of immunoblots represented by graph. GAPDH as a loading control. (C-D) Paraffin-embedded gastrocnemius (GAS) muscle sections (4μm) of WT (n=3), mdx (n=3) and mdx+TubA (n=5) mice processed for the detection of TUNEL positive nuclei (green). Sections were stained with α-laminin (red) to define the sarcolemma. Nuclei were counterstained with DAPI (blue). White boxed region shows the enlarged image from the mdx muscle section in the panel on right indicating TUNEL positive nuclei (white arrow). Numbers of TUNEL positive nuclei counted by Image J software. (E-F) Macrophage infiltration analyzed by Immunofluorescence staining of CD68 (green), α-laminin (red), and nuclei (blue) in GAS section isolated from WT (n=3), mdx (n=3), mdx+TubA (n=5) mice. White boxed region shows the enlarged image from the mdx muscle section indicating infiltration of CD68+ cells in the skeletal muscles (white arrow). Quantification of CD68 positive immune cells. Scale bar (50μm), Magnification-40X. *p < 0.05, †p < 0.01statistical difference between groups were determined using ANOVA with Tukey’s post hoc test.

Amelioration of dystropathology and improvement in muscle functional assessment following TubA treatment

Since pharmacological inhibition of HDAC6 improved MT acetylation, restored autophagic flux by promoting autophagosome-lysosome fusion, and decreased inflammation and apoptosis we next investigated whether TubA reduced muscle damage and improved muscle performance. Treatment of mdx mice with TubA prevented the increase in serum creatine phosphokinase (CPK activity (RI_TubA=133.81%, Figure 9A), a widely used clinical marker of muscle damage. The etiology of DMD can be explained by loss of membrane integrity due to dystrophin deficiency which results in degeneration of myofibers. To determine whether TubA affects sarcolemmal integrity, we intraperitoneally injected mice with Evans blue dye (EBD), which penetrates non-specifically into any cell with disrupted/leaky membranes. At 24 h after injection, EBD accumulated abundantly in mdx muscles as compared to WT muscles (Figure 9B), confirming membrane permeability due to the loss of dystrophin. Notable, TubA treatment significantly blunted EBD uptake into the DIA muscle of mdx mice. This observation was further confirmed by fluorescence staining of cross-sections of DIA muscles, the number of EBD positive fibers were reduced in mdx mice treated with TubA (RI _TubA=85.81%, Figure 9C).

Figure 9. HDAC6 inhibition ameliorates muscle pathophysiology and improves muscle function.

(A) Serum creatine phosphokinase quantitated by ELISA (n = 8 per group) (B-C) Loss of sarcolemmal integrity was evaluated by the i.p. Injection of Evans Blue Dye (EBD) in mice. Dye injected 24hr prior to the completion of two weeks TubA treatment. EBD staining of diaphragm (DIA) muscles and its cross section (4μm) showing EBD positive fibers. Quantification of EBD positive fibers in DIA muscle fibers (n=3 per group). Scale bar (90μm), Magnification-20X. (D) H&E-stained muscle (TA). White arrowhead= peripheral nuclei; black arrow head=central nuclei; black arrow=necrotic myofibers (infiltration of immune cells); black asterisk=fat depositions within myofibers. (n = 3 mice per group), Scale bar: 50 μm, Magnification-40X. Immunofluorescence staining of α-laminin (green) and DAPI (blue nuclei staining) in TA cross-section muscle of WT, mdx, TubA treated mdx. (n=3 per group) (E) Percentage fibers with centralized nuclei were quantified by Image J. Scale bar: 50 μm, Magnification-40X (F) Myofibers CSA calculated from minimum feret’s diameter (n=3 per group). (G-H) Grip strength, absolute and normalized to body weight (n = 5 per group). (I) Force–frequency relationship in DIA muscle strips from WT (n = 9), mdx (n = 8), and mdx+TubA (n = 9). (J). Eccentric contraction induced force drop (normalized to the first contraction) in EDL muscles isolated from WT (n=6), mdx (n=7), TubA treated mdx (n=6) mice. *p < 0.05,, ‡p < 0.001, statistical difference between groups were determined using ANOVA with Tukey’s post hoc test. *p<0.05 vs BL10, †p<0.01 vs BL10, ‡p<0.001 vs BL10, #p<0.05 vs mdx, ##p<0.01 vs mdx for panel I and J.

We next assessed the histopathological features of TA muscles from WT, mdx and TubA treated mdx mice by H& E staining. We observed severe necrosis (marked by black arrow), regenerating fibers with central nuclei (marked by black arrow head), and fat deposition in the interstitial space (marked by black asterisk) of dystrophic muscle as compared to WT muscles. Notably, overall necrotic myofibers, central nuclei, and fat depositions were reduced in TubA treated mdx (Figure 9D). We further quantified histological sections of TA muscles by immunostaining with α-laminin (green) and DAPI (blue) to analyze the cross-sectional area (CSA) and centronucleated myofibers in treated and untreated mdx mice (Figure 9D). A drastic increase in the percentage of fibers with centralized nuclei was observed in mdx TA muscle as compared to WT (46% versus 1.2%), which was significantly reduced in TubA-treated mdx muscle (RI_TubA= 52.2%, Figure 9E). TA muscle from mdx mice showed decreased cross-sectional area (CSA) as compared to WT, which was partially prevented with TubA treatment as exhibited by the distribution of Minimal Feret’s diameter (Figure 9F). DMD patients suffer from progressive muscle weakness, eventually leading to immobility and respiratory failure. To examine whether TubA improves muscle function and strength, in-vivo muscle functional performance was assessed by grip strength. Grip strength was conducted after completion of 2weeks of TubA treatment in mdx mice. Skeletal muscle strength was significantly improved in TubA treated mdx mice as compared to non-treated mdx mice (RI_TubA=113.29%, Figure 9G, H). Finally, to determine the effects on contractile function, we compared the force-generating capacities of untreated and TubA-treated mdx diaphragm muscle strips electrically stimulated ex vivo. TubA-treated mdx mice demonstrated significantly greater diaphragmatic force production over a broad range of stimulation frequencies (0 to 200 Hz) (Figure 9I, Table 1). In addition, TubA treated mdx EDL muscle was significantly protected from eccentric contraction induced force loss compared to mdx muscle (Figure 9J, Table 1). While TubA protected against contractile impairment we found no difference in passive stiffness of the diaphragm between mdx and TubA-treated mdx, 142.8 ± 52.5 and 126.7 ± 10.9 kPa, respectively.

Table 1.

Recovery Index for Tubastatin A treatment of mdx.

Outcome Measures	Recovery Score by TubA (70mg/kg b.w. for 2 weeks) in mdx mice
Serum Creatine phosphokinase (CPK)	133.8%
Immune Infiltration (CD68+ cells)	63.10%
Centralized nuclei	52.2%
Apoptotic cell death	92.14%
EDL ECC force production at 10th contraction	46.68%
Diaphragm Isometric force production (200Hz)	53.24%
Grip Strength normalized with Body weight	113.29%

Discussion

Previous studies have shown that defective autophagy and disorganized MT network play an important role in disease progression and contributes to DMD pathogenesis before the manifestation of severe phenotypes.^{11, 48–51.} Therefore, reverting autophagic dysfunction could be an efficient approach to slow the onset of disease progression and improve the muscle function in DMD patients. Our group was the first to identify that Nox2 mediates MT alterations and autophagic dysfunction in dystrophic muscle.^{7, 14} MTs regulate autophagy through their scaffolding and transport functions. Acetylated MTs induce autophagy by facilitating intracellular trafficking and fusion of autophagosomes with lysosomes.^{18, 19, 52} Despite the number of investigations, no consensus about the role of tubulin post translational modifications in regulating autophagic flux in DMD has been reached. We are the first to report that autophagy is differentially controlled by redox and acetylation modifications in mdx mice (Figure 10). We provide evidence that genetic ablation of Nox-2 in mdx mice promotes autophagosome maturation without facilitating autophagosome-lysosome fusion. Further, our study reveals alterations in MT acetylation in dystrophin-deficient muscle, which is not restored back in Nox-2 ablated mdx mice. Pharmacologically targeting HDAC6 restored MT acetylation, improved formation of the Q-SNARE complex to promote autolysosome formation, restricted muscle damage and improved muscle function in mdx mice.

Figure 10. Model of Impaired autophagy in mdx skeletal muscle are differentially regulated by redox and acetylation modifications.

(A) Loss of dystrophin leads to increased HDAC6 expression and subsequent deacetylation of α-tubulin. Decreased α-tubulin acetylation inhibits binding of the kinesin-/JIP-1 complex and subsequent transport of autophagosomes to lysosomes. Decreased activation of JIP-1 results in decreased phosphorylation of JNK and Bcl-2, preventing the dissociation of BECN1/Bcl-2 and sequestering BECN1 away from the PI3classIII complex (ATG14L-Vps34-Vps15). The net result is impairment of phagophore nucleation. Decreased WIPI-1 inhibits localization of ATG5–12/16L1 complex to the phagophore, leading to impaired autophagosome maturation in mdx mice. SNARE associated proteins play a major role in membrane-mediated events of autophagosome-lysosome fusion, another crucial step in autophagy process. In mdx mice, reduced interaction of SNAP29 with STX17 and VAMP8 leads to the inhibition of autophagosome-lysosome fusion. (B) Genetic inhibition of Nox-2 activity in mdx mice activates JIP-1, resulting in phosphorylation of JNK and Bcl-2, dissociation of BECN1 from Bcl-2 and induction of phagophore nucleation. In addition, Nox-2 inhibition promotes ATG5–12 complex formation and therefore autophagosome maturation. However, autophagosome-lysosome fusion is not improved upon inhibition of Nox-2 activity, likely due to no change in α-tubulin acetylation. Pharmacological inhibition of HDAC6 promotes α-tubulin acetylation and improves SNARE complex formation, thereby facilitating fusion of autophagosomes with lysosomes, improves autophagy, decreases dystropathology and improves skeletal muscle function. Increased acetylation of the antioxidant enzyme Prx II likely contributes to improved muscle function in mdx skeletal muscle. Figure created with BioRender.com

Emerging evidence suggests alterations in VPS15, VPS34 and BECN1 lead to impaired endosome and lysosomal maturation in Danon Disease and Glycogen Storage Disease (GSDII), and evocative of lysosomal myopathies.^{53, 54} Our data suggests disrupted PI3KclassIII complex in mdx muscle. Interaction of Beclin-ATG14L complex was reduced, whereas Beclin-Bcl-2 complex formation, which is known to inhibit autophagic induction, is increased in mdx muscle. We found decreased JNK activation, likely due to reduced levels of the scaffolding protein JIP-1, in mdx muscles and reduced phosphorylation of Bcl-2 (S70), thus inhibiting Beclin-ATG14L association for vesicle nucleation. In addition, WIPI-1, which recruits lipid phosphatidylinositol-3 phosphate (PI3P), mediates recruitment of the ATG5–12/16L1 complex and LC3 lipidation, was found to be reduced in skeletal muscle from mdx mice. Taken together, our data suggest that defects in autophagosome maturation is mediated by alterations in a Kinesin-JIP-JNK signaling pathway in dystrophic mice.

Recent studies have highlighted the role of STX17-SNAP29-VAMP8 Q-SNARE tertiary complex in promoting autophagosome-lysosome fusion.^{19, 55} We revealed that the blockage of autophagosome-lysosome fusion in mdx mice was not due to altered expressions of SNARE associated proteins but due to the inhibition of the interaction of STX17 and VAMP8 with SNAP29, which culminates in accumulation of autophagosomes.⁷

Previous work from our lab has shown that genetic ablation of Nox2-ROS improved autophagy in mdx mice. Nevertheless, the detailed mechanisms by which autophagy was rescued in the p47 KO/mdx mouse was not evaluated, which is required to improve the therapeutic outcomes for DMD patients. Our current data showed that p47 KO/mdx mice prevented the defects of autophagosome maturation by promoting vesicle nucleation and elongation of autophagosomes. However, our data revealed that genetic inhibition of Nox-2 was not able to promote the interaction between STX17-SNAP29 and VAMP8-SNAP29 and thus cannot promote autophagosome-lysosome fusion and content degradation in mdx mice. These findings may explain why we only observed partial rescue of dystrophic pathology in the p47 KO/mdx mouse.⁷

Based on the above findings, we speculate that autolysosome formation in mdx muscle is redox independent, leading us to explore the mechanisms that underlie the failure of delivery of autophagosomes and its fusion with lysosomes. Protein acetylation controls autolysosome formation and improves autophagic flux.^{19, 39, 56} Emerging evidence suggests that HDAC2 promotes autophagy through deacetylation of STX17, which increases its binding with SNAP29 and promotes the Q-SNARE complex formation.³⁹ Interestingly, our findings revealed the deacetylation of STX17 and increased HDAC2 in mdx mice, which suggests that reduced interaction of STX17-SNAP29-VAMP complex is independent of the STX17 acetylation and HDAC2.

Previous studies from our lab and others have reported dystrophin deficiency results in increased tubulin (α and β) content, disorganized MT lattice network and elevated tubulin modifications (detyrosinated α-tubulin, DT-tubulin) in mdx muscle.^{14, 57, 58} Post translational modifications of MTs, in particular acetylation, have been shown to facilitate autophagosome formation and serve to direct mature autophagosomes for fusion and degradation.^{41, 59, 60} In the present study, we observed decreased acetylated α- tubulin in young (5 weeks of age) mdx mice with increased protein expression of the deacetylase enzyme HDAC6. In a zebra fish model of DMD Pistocchi and colleagues⁶¹ also found decreased acetylated α-tubulin in skeletal muscle compared to non-dystrophic zebra fish and that inhibition of HDAC8 (class I HDAC) promoted α-tubulin acetylation and rescued muscle phenotype. On the other hand, Schaeffer and colleagues⁶² report increased acetylated α-tubulin in skeletal muscle from adult (11 weeks of age) mdx mice, with further increase upon HDAC6 inhibition and improved muscle phenotype. Young mdx mice, 2–8 weeks of age, undergo repetitive cycles of degeneration/regeneration while adult mdx mice, 10 weeks and older, the muscle has acquired a more stable phenotype and do not undergo cycles of degeneration/regeneration. Acetylation of α-tubulin has been shown to be increased at the late stage of myogenesis,^{63, 64} and therefore the increased acetylation of α-tubulin in adult muscle may be indicative of a more stable fully differentiated muscle phenotype. Importantly, inhibition of HDAC6 improved muscle phenotype in both young and adult mdx mice.

HDAC6 inhibition exerted beneficial effects by improving autophagic dysfunction through MT acetylation in several pathophysiological conditions such as Huntington’s disease (HD),²⁷ spinal cord injury,¹⁹ osteoarthritis,³³ and doxorubicin-induced cardiomyopathy.³⁰ HDAC6 has unique structure and properties, exerting both enzymatic and non-enzymatic effects on cellular functions. In addition to two catalytic domains which deacetylates cytoplasmic non-histone proteins including α-tubulin, peroxiredoxin, cortactin, β-catenin, heat shock protein (Hsp90), heat shock transcription facto-1 (HSF-1),^{65, 66} it also possesses a non-enzymatic zinc-finger ubiquitin-binding domain (UBD) at its C-terminus.⁶⁷ Through its UBD, HDAC6 interacts with ubiquitinated protein aggregates to promote loading and transport along the microtubules for proper degradation. HDAC6 has also been shown to regulate muscle atrophy through TGF-β and FoxO3 pathways.^{62, 68} Given this dual role of HDAC6 in regulating the response to cellular stress, inhibition of HDAC6 is a promising therapeutic strategy for several diseases. We selected TubA, which stands out as a potent and highly specific HDAC6 inhibitor with an IC₅₀ of 15 nM and has a strict selectivity for HDAC6 over all other HDACs (over 1000-fold) except for HDAC8 (57 folds).⁶⁹ Unlike other HDAC inhibitors, TubA exhibits no toxicities such as fatigue, nausea, or thrombocytopenia, thus making HDAC6 a most suitable and promising therapeutic target.⁶⁹ We assessed TubA dosing in mdx cohorts at 70mg/kg every day for 2 weeks. Our data showed improvement in muscle contractile properties and strength, and reduced serum CK levels in mdx mice. Moreover, HDAC6 inhibition specifically promoted tubulin acetylation without affecting DT-tubulin levels in mdx mice. Immunohistochemistry showed that the increased actylated-α-tubulin and α-tubulin are taking place within the muscle fibers themselves, confirming the changes observed by western blot. Intriguingly, HDAC6 inhibition did not restore the grid-like organization or the MT density of isolated FDB myofibers nor the passive stiffness of diaphgram strips in mdx mice. Interestingly, Schaeffer and colleagues⁶² found that HDAC6 inhibition increased the amount of transverse oriented α-tubulin in mdx mice. The difference between our studies and that of Shaeffer and colleagues may be due to the length of time mice were treated with tubastatin A (14 versus 30 days, respectively). Ward and colleagues have shown that selective inhibition of HDAC6 in isolated cardiomyocytes and FDBs from wild-type mice increased tubulin acetylation and cytoskeletal stiffness without altering the levels of α-tubulin and DT-tubulin.⁷⁰ Although Ward and colleagues did not quantify MT organization, their data suggest that HDAC6 inhibition does not alter the organization of the MT network in control wild-type mice. Taken together, our data along with that from Ward and colleagues highlight the importance of a proper balance of tubulin acetylation in order to maintain cellular signaling, analogous to redox balance.⁷¹ While improving tubulin acetylation under pathological situations such as mdx enhances autolysosome formation and improves muscle function, excessive acetylation in non-diseased muscle increases tissue stiffness and promotes Nox2-dependent ROS production.

Autophagosome maturation was not improved in TubA treated mdx as HDAC6 inhibition did not activate JIP/JNK signaling as assessed by phospho-JNK. HDAC6 inhibition was recently reported to restore neuronal autophagy flux and promoted the autophagosome-lysosome fusion by SNARE machinery in neurodegenerative diseases.^{19, 27, 72} Consistent with these findings, TubA treatment promoted Q-SNARE complex formation and improved autophagosome-lysosome fusion. We also found increased KIF5B expression in mdx skeletal muscle. KIF5B is a microtubule-based motor protein which transports mitochondria, lysosomes and myonuclei along microtubules. The upregulation of KIF5B may be compensatory, improving transport of autophagosomes and lysosomes to improve autophagy. KIF5B was found to be localized around central nuclei in both mdx and TubA-treated mdx muscle, which is to be expected as KIF5B plays an important role in positioning of myonuclei in skeletal muscle.^{73, 74} Altogether, HDAC6 inhibition facilitates autophagosome-lysosome fusion along MTs primarily by promoting tubulin acetylation rather than alterations in organization or density in mdx mice.

De Palma et al⁹ have shown that treating adult (16 weeks of age) mdx mice with the autophagy inhibitor chloroquine did not result in further accumulation of the autophagy markers LC3I, LC3II, and p62. Using the same dosing regimen in young (5 weeks of age) mdx mice we find that chloroquine further increased these autophagy markers. Given that dystrophin deficiency is a progressive disease these finding suggest that inhibition of autophagy is also progressive such that in the early stages of the disease (young, 5 weeks old mice) autophagy is not completely inhibited while in adult mice (16 weeks old) autophagy is severely inhibited. The levels of LC3I and p62, which are indicative of autophagy activity, decreased, indicating that there was an improvement in autophagic flux. However, it’s important to note that LC3II (the lipidated form of LC3) remained unaffected, suggesting that while there are improvements, TubA treatment may not completely restore autophagic flux. Lipidation of LC3 has been a hallmark of autophagy; however, there is increasing evidence that LC3 lipidation can also occur in cellular processes that are not directly related to autophagy (non-conical autophagy; reviewed in⁷⁵). Furthermore, lysosomotropic drugs such as chloroquine⁷⁶ as well as lysosomal damage⁷⁷ have been shown to promote LC3 lipidation independently from autophagy. Xu and colleagues⁷⁸ have shown lysosomal dysfunction in mdx skeletal muscle. This raised the possibility that the sustain elevation of LC3II observed in mdx mice is due to non-conical autophagy and/or lysosomal dysfunction, which is not reversed by TubA treatment.

HDAC6 may have a role in regulating redox balance within the cell, as HDAC6 has been shown to deacetylate redox regulatory proteins, PrxI and PrxII²⁵; and once deacetylated PrxI and PrxII lose their antioxidant capacity.^{24, 29, 32} We demonstrate that acetylation levels of PrxII were decreased in mdx muscle, and TubA treatment recovered both acetylation levels of PrxII and total PrxII in mdx mice. Our previous findings show that PrxII overexpression protects mdx mice from eccentric contraction induced force loss.⁴⁶ However, we did not observe any change in either oxidized PrxII (PrxSO2/3) or stretch dependent ROS production in mdx mice treated with TubA. Intriguingly, p47 KO/mdx mice showed reduced levels of oxidized PrxII (PrxSO2/3), increased PrxII acetylation, but did not increase MT acetylation or improve autolysosome fusion. Thus, there appears to be cross-talk between HDAC6 and Nox2 in dystrophic skeletal muscle that we have yet to understand. Future studies examining cross-talk between redox signaling and acetylation will allow development of effective therapeutics for maintaining muscle homeostasis in DMD patients.

Our data further reveals that TubA has the potential to act on the downstream pathogenic events of dystropathology. TubA treatment attenuated apoptosis, decreased levels of CD68+ macrophages infiltrated into the muscle, maintained sarcolemmal integrity, and remarkably reduced serum CK levels in mdx mice. In DMD patients, even at very young ages, weakness in hand strength is observed as a generalized effect of dystrophin deficiency. Loss of grip strength was observed in mdx mice while TubA treated mdx mice showed significant improvement in muscle strength. Moreover, TubA-treated mdx mice showed significantly greater diaphragm force production and decreased ECC induced force loss.

In conclusion, our study suggests that defects in autophagy are regulated by both redox and acetylation modifications in mdx mice. Evidence provided in this study confirms the role of HDAC6 in deacetylating MTs, which led to the disruption of autophagosome-lysosome fusion in dystrophin deficient mice. While genetic inhibition of Nox2 in mdx mice prevents MT disorganization and increased density,¹⁴ there is no effect on MT acetylation or autolysosme formation; suggesting that MT acetylation is the primary determinant of autophagosome lysosome transport and fusion in mdx mice. HDAC6 inhibition holds the potential to promote tubulin acetylation and improve autophagic flux in dystrophic skeletal muscle (Figure10). As HDAC6 has also been classified as a muscle atrogin, targeting HDAC6 serves as an effective tool in preventing disease progression and improving muscle strength and function (Table 1), supporting the translational potential of TubA into clinical research for DMD patients.

Material and Methods

Antibodies and Reagents

Tubastatin A and PEG300 were purchased from Selleck Chem, DCFH-DA (6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate) was from Invitrogen. DMSO, Chloroquine and Evans Blue Dye were purchased from Sigma Aldrich. Tween-80 from Fischer Scientific. Protein A Magnetic Beads, Mini-Protean TGX stain free gels, and Immuno-Blot^® PVDF Membrane, Clarity western ECL substrate were purchased from Bio-Rad. Anti-beclin, anti-p-bcl-2, anti-vps34, anti-vps15, anti-JNK, anti-p-JNK, anti-atg5–12, anti-atg7, anti-atg4b, anti-α-tubulin, anti-HDAC6, anti-p62, anti-caspase-3, and anti-LC3 antibodies were from Cell Signaling Technology. Anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) and anti-dynein were purchased from EMD Millipore. Anti-CD68, anti-LAMP2, anti-Cathepsin-B, anti-SNAP29 were purchased from Santa Cruz Biotechnologies. Anti-WIPI1, anti-snap29, anti-stx17, anti-vamp8, anti-KIF5B, anti-KLC, anti-detyrosinated, anti-Prx-SO3, and anti-α-laminin were from Abcam. Anti-acetylated tubulin, anti-bcl-2, anti-atg14, anti-MAPKK, anti-JIP-1, anti-prxII were from Millipore Sigma. Secondary antibodies for immunofluorescence (Alexa Fluor^® 594 donkey anti-rabbit, Alexa Fluor^® 488-donkey anti-rabbit, Alexa Fluor^®594-goat anti-mouse), and ProLong^™ Gold Antifade Mountant with DAPI were purchased from Invitrogen. Microscopic slides was purchased from Denville Scientific Inc. For detailed information about antibodies and dilution see Supplementary Table 1.

Mice

C57BL/10ScSnJ (WT) and C57Bl/10ScSn-Dmdmdx/J (mdx) were purchased from Jackson Laboratories (Bar Harbor, ME) and bred following their breeding strategy. Mice with a genetic deletion of the Nox-2 scaffolding subunit p47^phox (p47 KO) were obtained from The Jackson Laboratory (B6 (Cg)-Ncf1m1J/J) and bred onto the mdx background as previously described.⁷ All animals used for experiments were males between 3–6 weeks of age. All animal studies were conducted in accordance with ARRIVE (Animal Research: Reporting of In Vivo Experiments). Mice were housed in a specific pathogen free (SPF) facility on a 12hr light/dark cycle. All mice were monitored daily and drug tolerability evaluated on the basis of body weight and clinical signs. After treatment period, mice were sacrificed by deep anesthesia followed by cervical dislocation as approved by the Institutional Animal Care and Use Committee (IACUC) of Baylor College of Medicine. Skeletal muscles were quickly excised, and either placed in Krebs-Ringer buffer for tissue experiments, snap frozen in liquid nitrogen for biochemical and molecular experiments, or in 10% formalin for histological experiments. We employed basic methods to analyze the preclinical experiments (TREAT-NMD).⁷⁹

Autophagy Inhibitor

To explore the role of autophagy, autophagy inhibitor, chloroquine (CQ)(Sigma) was selected. CQ was dissolved in MilliQ with a final concentration of 50mg/ml. CQ was intraperitoneally (i.p.) injected into 3 week-old mdx mice (onset of disease progression) at the dosage of 50 mg/kg b.w. once a day for 07days.⁹

HDAC6 Inhibitor treatment

For pharmacological inhibition of HDAC6, HDAC6-Specific inhibitor TubastatinA (TubA) (Selleck-Chem, USA) was dissolved in 4% dimethyl sulfoxide (DMSO), 30% PEG300 and 66% MilliQ with a final concentration of 2.5 mg/mL, and was freshly prepared each day. TubA was intraperitoneally (i.p.) injected into 3 week-old mdx mice (onset of disease progression) at the dosage of 70 mg/kg b.w. once a day for 2 weeks.

Fore-limb Grip strength analysis

Fore-limb grip strength was performed to assess the effects of TubA on recovery of muscle strength of mdx mice. The grip strength meter (Columbus Instruments, OH) was positioned horizontally and mice were held by the tail. Mice were allowed to grasp the smooth metal rod with their forelimbs only and then were pulled backward by the tail. The force applied to the metal rod at the moment the grasp was released and recorded as the peak tension. The test was repeated 5 consecutive times within the same session and the average value from the 5 trials was recorded as the grip strength for that animal. In order to avoid modifications induced by the body weight of the mice in the latency of staying on the grip strength meter, the data have been normalized with body weight of mice. The grip strength test was conducted by the same investigator in order to avoid variability in performing experiment, and the investigator was blinded to the treatment group.

Serum Collection

Blood was collected from the hepatic portal vein of mice immediately following sacrifice and left at room temperature for 30 min to achieve coagulation. Serum was then separated from other blood fractions by centrifugation at 1,000 g for 10 min and stored at −80 °C for further use. CK activity in the serum was measured using Creatine Kinase Activity Assay Kit (My BioSource) following the manufacturer’s instruction.

Ex-Vivo EDL Eccentric force measurements

Muscle contractile force measurement was conducted by using an ex-vivo muscle test system. In brief, the mice were anesthetized with isoflurane, the hind limb was excised and immediately placed in a bicarbonate-buffered solution (120 mM NaCl, 4 mM KCl, 1 mM MgSO₄, 1 mM NaH₂PO₄, 25 mM NaHCO₃, and 2 mM CaCl₂, 10mM glucose) equilibrated with 95% O₂-5% CO₂ (pH7.4) for dissection. The proximal and distal tendons were tied with braided silk suture thread (4–0, Fine Science Tools) and mounted in a muscle bath containing bicarbonate-buffered solution continuously bubble with 95%O₂, 5% CO₂ between a fixed hook and a dual-mode lever system (305C-LR-FP, Aurora Scientific Inc., Aurora, ON, Canada) and allowed to equilibrate to 30°C for 10 minutes. The stimulation protocol consisted of supramaximal electrical current delivered through platinum electrodes using a biphasic high-power stimulator (701C; Aurora Scientific). Optimum length (Lo) was determined with a series of twitch stimulations measured using a hand-held electronic caliper, after which the muscle underwent 10 eccentric contraction with 3 minutes rest between each contraction so as to not elicit muscle fatigue. Each eccentric contraction consisted of a 200 ms isometric tetanus (150Hz) followed by stretch from 100% to 110% of Lo at 0.5Lo/s and then shortened to Lo passively. Following the 10 eccentric contractions the muscle was removed from the organ bath, trimmed of connective tissue, blotted dry, and weighed. Data were analyzed using the dynamic muscle control and analysis software (Aurora Scientific Inc.).

Ex-Vivo Diaphragm force measurements

Diaphragm (DIA) muscle was dissected from mice and one end tied to a fixed hook and the other to a force transducer (F30, Harvard Apparatus) using silk suture (4–0) in bicarbonate-buffered solution continuously gassed with 95% O₂–5% CO₂ at 30°C. Contractile properties were assessed by passing a current between two platinum electrodes at supramaximal voltage (PanLab LE 12406, Harvard Apparatus) with pulse and train durations of 0.5 and 250 ms, respectively. Muscle length was adjusted to elicit maximum twitch force (optimal length, Lo) and the muscle was allowed a 10-min equilibration period. To define the force-frequency characteristics, force was measured at stimulation frequencies of 1, 5, 10, 20, 40, 60, 80, 120, 150 and 200 Hz every 1 min. At the end of the contractile protocol, muscle length was measured using a hand-held electronic caliper, fiber bundles removed from the organ bath and trimmed of excess bone and connective tissue, blotted dry and weighed. Muscle weight and Lo were used to calculate normalized forces expressed as N/cm². Diaphragm passive stiffness was determined during passive stretch of the diaphragm as we have previously described.¹⁴

ROS measurements

Diaphragm intracellular ROS was measured using 6-carboxy-20,70-dichlorodihydrofluorescein diacetate (DCFH-DA) (Invitrogen, Carlsbad, CA) as previously described.¹⁴ Briefly, diaphragm muscle optimal length was determined as described above followed by incubation with DCFH-DA for 30 min, washed using the physiological saline solution and de-esterified for an additional 30 min at 25°C. All cell-loading and imaging was performed in the dark to prevent light induced oxidation of DCFH-DA. DCF was excited at 470/20 nm using a Sutter Lamda DG-5 Ultra-high-speed wavelength switcher and emission intensity was collected at 535/48 nm on a charge coupled device (CCD) Camera (CoolSNAP MYO, Photometrics, Tucson, AZ) attached to an Axio Observer (Zeiss) inverted microscope (20X objective, 0.5 NA) at a rate of 0.2 Hz. Alterations in the rate of ROS production were baseline corrected and calculated over the final minute of the stretch period.

Western blotting and Immunoprecipitation

TA muscle from hind limb of the mice were isolated and lysed with ice-cold RIPA buffer containing protease and phosphatase inhibitors, and centrifuged at 13,500rpm for 10 min. The supernatant was retained, aliquoted, and the protein content was quantified using the bicinchoninic acid (BCA) Assay Kit. A volume corresponding to 10 μg of protein was diluted with a Laemmli sample buffer (BioRad), and heated at 100 °C for 5 min. Protein samples were separated via 4–15% and 4–20% SDS-PAGE and then transferred to polyvinylidene difluoride (PVDF) membranes, which were blocked with 5% non-fat milk or 5% BSA (dissolved in Tris-buffered saline, pH 7.4, and 0.2% Tween 20, TBST) for 60 min at room temperature.

For immunoprecipitation, total protein lysates were prepared and quantified as described above. Protein lysates (1000μg) were incubated with protein A magnetic beads (BioRad) as per the manufacturer’s instructions. Protein complexes were analyzed by western blotting using indicated antibodies (Beclin-1, ATG14L, Bcl-2, SNAP29, VAMP8, STX17, PrxII, acetyl lysine). Immunoblots were incubated overnight at 4 °C with primary antibodies. The following day, membranes were washed and incubated with secondary antibodies for 60min at room temperature. Clarity western ECL substrate (Bio-Rad) was added on the membrane and the signals were visualized by ChemiDoc Imaging system (Bio-Rad). The bands were quantified using densitometric analysis by the ImageLab Software.

Single Fiber isolation from FDBs

For α-tubulin staining, FDB fibers isolated from mice and incubated in DMEM (ThermoFisher Scientific) containing 0.1% penicillin-streptomycin (ThermoFisher Scientific) and 0.4% Collagenase A (Sigma) at 37 °C for 2.0 h. Following enzymatic digestions, isolated fibers were transferred to DMEM media with 10% FBS for overnight. The following day, isolated fibers were fixed in 10% formalin for 20 min at RT. The fibers were rinsed 3 times in 1X PBS with 1mM EGTA and further used for immunofluorescence staining(α-tubulin conjugated with Alexa Fluor 488) and fibers were imaged using Zeiss LSM 880 with Airyscan FAST confocal microscope and directionality and density assessed as we have previously described.¹⁴

Histology and Immunofluorescence

The skeletal muscles (TA and GAS) were carefully excised from hind limb of mice and post-fixed in 10%formalin overnight. Paraffin blocks were then prepared after dehydration, clearing, and wax impregnation. Transverse Sections (T.S.) of 4μm thickness were cut with a rotary microtome, deparaffinized in xylene and gradient of alcohol were subjected to histological or immunofluorescence staining.

For H&E staining, the slides were stained in hematoxylin solution for 20–30 min and rinsed in running water three times. Slides were then stained in eosin solution for 1–2 min, dehydrated in graded ethanol and xylene, and covered using DPX paramount for further imaging. For immunofluorescence, serial muscle cross-sections (4μm) were cut from each paraffin block by rotatory microtome. The sections were placed on coated slides (Denville Scientific Inc.) and dried at 60 °C in a hot air oven. Sections were deparaffinized, rehydrated, and underwent an antigen retrieval method using citrate-EDTA solution incubated at 96 °C for 20 min in water bath. After dipping the slide in distilled water, the sections were blocked with 3% BSA for 60min at room temperature, and incubated with the primary antibodies overnight at 4°C followed by secondary antibodies at room temperature for 1 h. Primary antibodies include acetylated α-tubulin, α-tubulin, SNAP29, VAMP8, CD68, α-laminin, KIF5B, and PrxII. Secondary antibodies include Alexa Fluor 488 and 594. Tissue sections were mounted with Prolonged Gold Antifade with DAPI (Invitrogen). All images were taken with fluorescent microscope (ECHO, CA) equipped with a 40× 0.75NA objective. Alexa Fluor 488 was excited at 470/40 and emission collected at 525/50 while Alexa Flour 594 was excited at 580/30 and emission collected at 635/50 in sequential images so as to prevent crosstalk between the fluorescent channels. Images were background subtracted and co-localization analysis was carried out by ImageJ software.

For CD68 quantification, images acquired from five different optical fields of GAS muscle cross-section (40X magnification) and the number of CD68+ immune cells were counted per area (μm²).

The CSA and centralized nuclei were measured based on α-laminin staining and DAPI stained nuclei respectively. CSA was quantified by minimum Feret’s diameter and central nucleation was quantified as % fibers with central nuclei. TA muscle cross-sections were used for quantification. All the immunofluorescence images were quantified using Image J software (NIH, USA).

Colocalization analysis

For double immunofluorescence, two primary antibodies (SNAP29 and VAMP8) were incubated on the same gastrocnemius (GAS) tissue sections. Three serial transverse section of the GAS muscle were analyzed from 3–5 different optical field per animal. To quantify the degree of colocalization between the anti- SNAP29 and anti-VAMP8 staining, Pearson’s correlation coefficient was performed using ImageJ Software.

RNA extraction and Gene expression analysis

Total RNAs were extracted from gastrocnemius (GAS) muscles using TRIzol reagent (Invitrogen) according to the manufacturer’s protocol. Gene expression was verified by using iScript cDNA synthesis kit (Biorad) for reverse transcription and Real Mastermix for PCR following manufacturer’s instructions. The primers used were as follows:

• BECN1-FWD (5’ AGG CTG AGG CGG AGA GAT T-3’),

• BECN1-REV (5’- TCC ACA CTC TTG AGT TCG TCA T-3’);

• ATG5-FWD (5’-ATCAGACCACGACGGAGCGG-3’),

• ATG5-REV (5’-GGCGACTGCGGAAGGACAGA-3’);

• ATG12-FWD (5’- ACAAAGAAATGGGCTGTGGAGC-3’),

• ATG12-REV (5’- GCAGTAATGCAGGACCAGTTTACC-3’);

• ATG14L-FWD (5’- GCAGCTCGTCAACATTGTGT-3’),

• ATG14L-REV (5’-TGCGTTCAGTTTCCTCACTG-3’);

• VPS34-FWD(5’-TGTCAGATGAGGAGGCTGTG-3’),

• VPS34-REV(5’-CCAGGCACGACGTAACTTCT-3’);

• GABARAPL1-FWD (5’- CGGTCATCGTGGAGAAGGCT-3’),

• GABARAPL1-REV (5’- CCAGAACAGTATAACGGCAACTCC-3’);

• FIP200-FWD(5’-ACCGTGCACCTGCTATTCCT-3’),

• FIP200-REV (5’ CATCATGGACAAGCCCTTCA-3’);

• UVRAG-FWD (5’ CAA GCT GAC AGA AAA GGA GCG AG-3’),

• UVRAG-REV (5’- GGA AGA GTT TGC CTC AAG TCT GG-3’);

• HDAC6-FWD(5’-AAGTGGAAGAAGCCGTGCTA-3’),

• HDAC6-REV (5’- CTCCAGGTGACACATGATGC −3’)’;

• MEC17/ATAT1-FWD (5’- ACTGA AGGAC ACCTC AGCCC GA −3’),

• MEC17/ATAT1-REV (5’- TACCT CATTG TGAGC CTCCC GG-3’)

Evan’s Blue Dye uptake

To assess the membrane integrity of WT, mdx, and TubA treated mdx mice, Evans Blue Dye (EBD) was injected to the animal as described by Millay et al.⁸⁰ EBD (10mg/ml) was dissolved in PBS (pH 7.4) and sterilized by passage through membrane filters with a 0.2-mm pore size. Mice were injected intraperitoneal (i.p.), with 0.1 ml/10g body weight with EBD 24h prior to the completion of two week treatment with TubA. After 24h, DIA muscles were harvested from all the mice groups and sectioned using rotatory microtome. EBD positive fibers showed a bright red emission under fluorescent microscope. All sections were examined, photographed and muscle fiber counted under a fluorescence microscope (ECHO, CA).

TUNEL ASSAY

GAS isolated from hind limb of mice fixed in 4% formalin overnight and then paraffin-embedded muscles sections (4μm) were assessed by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) assay with an in-situ Cell Death Detection Kit (Biovision Inc., CA, USA) as per the manufacturer’s instructions. Antigen retrieval was performed, and slides were then blocked in 3% bovine serum albumin for 1hr at RT and incubated in anti-α-laminin overnight at 4°C. The following day, slides were incubated in donkey anti-rabbit AF594 and then stained with Prolong gold antifade reagent with DAPI (Invitrogen). Tissue sections were examined under fluorescence microscope (ECHO, CA). TUNEL positive nuclei were counted manually as a co-localization of TUNEL + nuclei with DAPI and expressed as a percentage of the TUNEL + nuclei per total number of DAPI stained nuclei obtained from five different optical field per muscle section (40X Magnification). Total number of DAPI stained nuclei were counted using Image J software (NIH).

Statistical analysis

Statistical analysis were performed using Origin Pro Software (OriginLab Corporation, Northhampton, MA, USA). Differences between two groups were analyzed using Student’s t-tests and normality tests. Differences between means of multiple groups were analyzed by one or two-way analysis of variance ANOVA followed with Tukey post-hoc tests. Values of p<0.05 (95% confidence) were considered statistically significant. Data are represented by box plots with mean (empty checker box within the box), median (solid bar), SD (whisker), SE (box). The number of mice used in groups for each parameter are indicated in figure legends. Recovery index (RI_TubA) was calculated as ((mdx+TubA)-(mdx))/((WT)-(mdx))x100.⁷⁹

Data Availability Statement:

The data that supports the findings of this study are available in the online supplementary material of this article.