Botulinum toxin A-induced muscle paralysis stimulates Hdac4 and differential miRNA expression

At sufficient dose, intramuscular injection of Botulinum toxin A causes muscle wasting that is physiologically consistent with surgical denervation and other types of neuromuscular dysfunction. The aim of this study was to clarify early molecular and micro-RNA alterations in skeletal muscle following Botulinum toxin A-induced muscle paralysis. Quadriceps were analyzed for changes in expression of micro- and messenger RNA and protein levels after a single injection of 0.4, 2 or 4U Botulinum toxin A (/100g body weight). After injection with 2.0U Botulinum toxin A, quadriceps exhibited significant reduction in muscle weight and increased levels of ubiquitin ligase proteins at 7, 14 and 28 days. Muscle miR-1 and miR-133a/b levels were decreased at these time points, whereas a dose-responsive increase in miR-206 expression at day 14 was observed. Expression of the miR-133a/b target genes RhoA, Tgfb1 and Ctfg, and the miR-1/206 target genes Igf-1 and Hdac4, were upregulated at 28 days after Botulinum toxin A injection. Increased levels of Hdac4 protein were observed after injection, consistent with anticipated expression changes in direct and indirect Hdac4 target genes, such as Myog. Our results suggest Botulinum toxin A-induced denervation of muscle shares molecular characteristics with surgical denervation and other types of neuromuscular dysfunction, and implicates miR-133/Tgf-β1/Ctfg and miR-1/Hdac4/Myog signaling during the resultant muscle atrophy.


Introduction
In addition to cosmetic applications, intramuscular injection of Botulinum toxin A (BoNT/A) has been used to treat a range of conditions with underlying muscle dystonia, including facial palsy, pain and muscle spasticity [1][2][3][4]. This neurotoxin acts by targeting SNARE protein synaptosomal-associated protein 25 (SNAP-25) for cleavage, which prevents the regulated secretion of neurotransmitter acetylcholine (Ach) at the neuromuscular junction (NMJ). Inhibition is transient, lasting about 3-4 months in humans and about 4 weeks in mice [3], and recovery involves synaptic remodeling of the NMJ. However, at sufficient dose, BoNT/A causes rapid muscle atrophy, a response that is conserved across species [5,6]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Animals and BoNT/A-induced denervation model
All animal experiments were performed using protocols and procedures approved by the Institutional Animal Care and Use Committee of the University of Washington. Female C57Bl/6 mice were obtained from Jackson Laboratories (Sacramento, CA, USA) and were 20-26 weeks of age at the initiation of the experiments. Transient muscle paralysis was induced in the right quadriceps muscle group via a single injection of botulinum neurotoxin A (BoNT/A) as described previously [25]. Quadriceps paralysis was confirmed 24 hours post-injection by visual examination of reduced extension of the affected limb [25,26]. Animals were allowed free cage activity and food and water ad libitum throughout the experiment.
For miRNA/mRNA analyses, mice were randomized into three BoNT/A dosing groups (Units of BoNT/A/100g body weight) of 0.4U, 2.0U and 4.0U (n = 4-6/dosage), or untreated age-matched controls (Naïve, n = 10). For protein analyses, mice were randomized into a 2.0U BoNT/A dosage group and Naïve controls. This arm of the experiment was limited to a single BoNT/A dosage to minimize animal use. Mice from each group were then chosen at random for euthanasia on day 7, 14 or 28 and specimen collection as described below (n = 4-6/group/ time point in BoNT/A, n = 10 Naïve controls). Day 28 was chosen as the final time point, as this was the point of maximal BoNT/A-induced muscle atrophy in our previous studies [26].

RNA preparation and quantitative RT-PCR
Quadriceps muscles were collected into RNAlater Solution (Ambion) and stored at 4˚C until all experimental samples were collected (up to 28 days). A 25mg section of the mid belly of the quadriceps was cut for further processing. miRNA and mRNA were extracted from this muscle section using the miRNeasy Mini and RNeasy MinElute Cleanup kits (Qiagen). miRNA was analyzed using TaqMan Small RNA Assays (Applied Biosystems: catalog # 002222, 002246, 002247, 000510, 001973) and ViiA7 Real-Time PCR System (Applied Biosystems; Thermo Fisher). cDNA was synthesized from mRNA using Superscript III reverse transcriptase (Thermo Fisher), and analyzed by quantitative RT-PCR using SYBR green on the ViiA7. Gene expression levels relative to U6 small nuclear RNA and β-actin for miRNA and mRNA, respectively were quantitated using the 2^(-ΔΔCT) method. Canonical muscle miR-1, -133a, -133b and -206 were investigated. Validated target genes of these muscle specific miRNAs (shown in Table 1) were chosen due to their reported roles in muscle. Primer sequences used in this study are listed in S1 Table.

Immunoblotting and antibodies
Left and right quadriceps muscles were weighed and then snap frozen in liquid nitrogen. Samples were ground to a powder in liquid nitrogen using a mortar and pestle, collected into RIPA lysis buffer (50mM Tris-HCl, pH8, 150mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) and incubated on a rotating platform at 4˚C for 1h. Lysate was clarified by centrifugation at 16000×g for 10 min and supernatant quantitated by BCA assay (Thermo Scientific). Samples were run on 4-12% NuPAGE gels (Thermo Scientific) and transferred to PVDF membranes (Bio-Rad). Membranes were probed with antibodies against Atrogin-1 (MAFbx; sc-166806, Santa Cruz Biotechnology), MuRF-1 (sc-398608, Santa Cruz Biotechnology) or Hdac4 (#7628, Cell Signaling Technology), with β-actin (ab8227, Abcam) as a loading control. Primary antibodies were visualized by chemiluminescence (Thermo Scientific) after incubation with HRPconjugated secondary antibodies (GE Healthcare or Santa Cruz Biotechnology).

BoNT/A induces muscle atrophy and activates myogenic gene expression
We previously reported maximal calf muscle atrophy at 28 days after a single BoNT/A injection [26]. Here we observed significant reduction in muscle weight at 7, 14 and 28 days post BoNT/A treatment, compared with naïve quadriceps (Fig 1A). There was no significant difference in weight of injected muscles between day 14 and 28 (quadriceps masses were 47±1 and 48±6% of naive values, respectively). Upregulation of ubiquitin ligases Atrogin-1 and MuRF-1 in BoNT/A injected muscles was confirmed by Western blot (Fig 1B). At day 14, we observed a significant increase of Forkhead box O1 (Foxo1) transcription factor, one of several factors that mediate upregulation of atrophy related ubiquitin ligases (Fig 1C; [59,60]).
Previous observations of myogenic gene expression following BoNT/A injection have varied, possibly due to differences in injection protocols and muscle groups studied [7,45,61,62]. We did not observe significantly altered expression of nAChRs Chrna1 and Chrng which previously were found to be regulated with denervation and BoNT/A-induced skeletal muscle atrophy [63]; however, there was an increase of the early response gene Musk at days 7 and 14 Igf-1 (insulin like growth factor 1) Growth factor targeted during myocardial infarction; inhibits cardiomyocyte apoptosis [48,49] Met (met proto-oncogene) Tyrosine protein kinase upregulated in rhabdomyosarcoma [50] Utrn (utrophin) Cytoskeletal protein downregulated during skeletal muscle differentiation [51] miR-133a/b Srf (serum response factor)

Transcription factor involved in myoblast proliferation and myogenesis [39]
RhoA (ras homolog family member A) Rho GTPase involved in cardiac hypertrophy [52,53] Tgfb1 (transforming growth factor beta 1) Growth factor upregulated in atrial fibrosis [54] Ctgf (connective tissue growth factor) Growth factor targeted in bladder smooth muscle cell fibrosis; fibrosis of cardiac myocytes [55,56] Egfr (epidermal growth factor receptor) Growth factor receptor targeted in prostate cancer cells [57] Gja1 (gap junction protein, alpha 1) Gap junction protein regulated during heart regeneration in zebrafish [58] https://doi.org/10.1371/journal.pone.0207354.t001 post injection (Table 2). There were also increased expression of genes encoding myogenic regulatory factors Myod, Myog and Myf5 at later time points, and enhanced expression of collagen component genes Col1a1 and Col3a1 at 28 days compared to the naïve control.

BoNT/A-induced muscle paralysis differentially alters expression of muscle miRNA
Muscle miRNA expression is altered by a variety of muscle atrophy pathologies, including denervation [33,64]. Here we also observed altered expression of muscle miRNAs following BoNT/A-induced muscle paralysis. The bicistronic partner transcripts miR-1 (Fig 2A) and miR-133a ( Fig 2B) demonstrated decreased expression at all BoNT/A doses and time points  compared to naïve mice (ranging from -32 to -86% and -60 to -81% respectively). miR-133b expression ( Fig 2C) was also decreased compared to naïve mice, with the exception of the lowest dose at the earliest time point. In contrast, miR-206 (Fig 2D), which shares a genomic locus with miR-133b, exhibited increased expression at the 2.0 and 4.0U BoNT/A doses at the later collection days (14 and 21) compared to naïve mice (6-to 12-fold elevation). We observed a dose related increase in miR-206 expression at day 14 but not at other time points. Finally, while the non-muscle specific miR-29b has a role in some models of muscle atrophy [65], here expression was 3 orders of magnitude lower than the examined muscle miRNAs, and was not altered by paralysis (S1 Fig).

BoNT/A-induced muscle paralysis activates known miR-133a/b target genes
Contradictory roles for miR-133 in targeting expression for the promotion of myoblast proliferation, and the repression of myoblast proliferation to drive differentiation [39,44], suggest that the role of these miR-133 sequences may be context specific. Levels of expression of validated miR-133 target genes were thus investigated after injection of quadriceps with 2.0U BoNT/A. While, miR-133 has been found to regulate Srf expression in muscle during proliferation [39], we did not observe significantly altered Srf mRNA expression at any time point after BoNT/A injection (Fig 3A). For the GDP-GTP exchange protein RhoA, levels of mRNA were significantly increased at day 14 and 28 following muscle paralysis vs naïve quadriceps (7-and 9-fold elevation, respectively; Fig 3B). Increased expression was also observed for Tgfb1 at the same time points (>30-fold elevation; Fig 3C). Another growth factor, Ctgf, regulated by miR-133 in muscle cells [55,56], showed a 15-fold increase in mRNA levels at day 28 after BoNT/A injection compared to the naïve control ( Fig 3D). Likewise, there were higher levels of Egfr at this timepoint (5-fold increase; Fig 3E). Finally, expression of Gja1, a target of miR-133 [58], was significantly elevated at day 14 and 28 (3-and 6-fold respectively; Fig 3F). Interestingly, Gja1 was previously also found to be a target of miR-1/206 in several mammalian contexts [66,67].

BoNT/A-induced paralysis has differential effects on miR-1/206 target gene expression
As miR-1 and miR-206 share a common seed sequence, they have been reported to affect many of the same mRNA target sequences. For example, Pax3 and Pax7, transcription factors that are involved in the transition between proliferative and differentiating satellite cells [68], are both targeted by miR-1 and miR-206 [46,47]. While we did not observe altered Pax3 mRNA expression (Fig 4A), Pax7 expression was elevated 5-fold at 14 days after paralysis ( Fig 4B). Hdac4, a transcriptional repressor of muscle genes, has also been found to be downregulated in muscle by both miR-1 and miR-206 to enable myogenic differentiation [32,39]. However, in this study Hdac4 expression was elevated relative to naïve muscles and expression reached statistical significance at day 28 following paralysis (60-fold; Fig 4C). For the proto-oncogene Met, there was a 14-fold increase in expression at day 14 after BoNT/A injection compared to the naïve controls ( Fig 4D). There were also increased levels of Igf-1 at day 14 and 28 after paralysis (8-and 10-fold elevation vs. naïve; Fig 4E). For miR-206 target, Utrn, there were no significant changes in expression after BoNT/A injection (Fig 4F).

BoNT/A-induced paralysis increases Hdac4 protein and alters Hdac4 target genes
Hdac4 is an inhibitor of muscle differentiation [69] and increased levels of this protein have been observed after surgical denervation and in models of muscle atrophy associated with amyotrophic lateral sclerosis (ALS) [12,13]. Consistent with these findings, increased Hdac4 protein following BoNT/A-induced muscle paralysis was detected in the present study by Western blot (Fig 5A). We then assessed expression of the Myog inhibitors Dach2 (dachshund 2) and Hdac9, which are known to be downregulated by Hdac4 during surgical denervation [12,13]. Dach2 expression, low in naïve muscles, was significantly lower in paralyzed muscles at all time points after BoNT/A injection (-80%; Fig 5B). Hdac9 expression was found to be elevated at day 28 post paralysis compared to naïve muscles ( Fig 5C). Finally, we observed that expression of Fgfbp1, a direct downstream target of Hdac4 in muscle after sciatic nerve crush [32], was reduced in paralyzed quadriceps at all days tested compared to naïve muscle (ranging from -70% to -80%; Fig 5D).

Discussion
We investigated early molecular responses in muscle following a single intramuscular injection of BoNT/A. BoNT/A-induced muscle paralysis caused rapid and significant muscle atrophy with concurrent changes in expression of key elements of proteasomal protein degradation. At a BoNT/A dose previously associated with severe muscle atrophy (2.0U/100g), we, for the first time, report significantly altered expression of muscle specific miRNAs. Muscle paralysis resulted in significant upregulation of Hdac4 mRNA and protein similar to reports of paralysis following surgical denervation [12,13]. Taken together, our data suggest that transient muscle paralysis induces altered gene expression that is similar to expression profiles associated with denervation and a variety of neuromuscular pathologies. With respect to miRNA expression, BoNT/A-induced muscle paralysis precipitated similar alterations as previously observed in alternate neuromuscular atrophy models. Expression of miR-1, miR-133a and miR-133b were all reduced following BoNT/A injection, while miR-206 was significantly elevated at 14 and 28 days post paralysis. This pattern of altered muscle miRNA expression recapitulates that observed following sciatic nerve transection [33,70]. Additionally, increased miR-206 levels have also been associated with onset of neurological symptoms in atrophy models of ALS and SMA (spinal muscular atrophy) [32,34].
miR-133a and miR-133b expression were decreased in response to BoNT/A-induced muscle paralysis, with subsequent upregulation of miR-133 target genes previously implicated in Tgf-β1 mediated hypertrophy and fibrosis of other tissues. For example, in a model of atrial fibrillation, nicotine treated atrial fibroblasts demonstrated reduced levels of miR-133, increased expression of Tgfb1 and its downstream target Ctgf, as well as higher levels of collagen production and atrial fibrosis [54]. In a rat model of bladder outlet obstruction, miR-133 downregulation in bladder smooth muscle cells resulted in similar Tgf-β1 profibrotic signaling, with accumulation of ECM and growth factors consistent with bladder wall hypertrophy and fibrosis [55]. In cardiac myocytes, decreased miR-133 levels during pathological hypertrophy led to increased Ctgf and collagen synthesis and resultant heart failure [56], and was also associated with increased levels of RhoA, which inhibits axon growth and regeneration [71], and is involved in cardiac hypertrophy [52,53]. Our observed upregulation of ECM genes are consistent with altered expression 4 weeks post-BoNT/A injection that were identified in a transcript profiling study [7]. It is interesting to speculate that the ECM remodeling and muscle fibrosis underlying BoNT/A-induced changes in muscle mechanical properties [5,16] may be modulated by downregulation of miR-133a/b and a consequent upregulation of ECM genes and growth factors, as observed here.
Based on the literature, we hypothesized a role for miR-1 in the proteasomal degradation that likely results in NMJ degeneration following BoNT/A-induced muscle paralysis [9, 15]. BoNT/A injection led to differential downregulation of miR-1 and dose-and time-dependent upregulation of miR-206, similar to muscle miRNA expression changes induced by surgical denervation [32,33,70]. We also detected upregulation of the downstream targets Pax7, Met, Hdac4 and Igf-1. As expression of the miR-206 target Utrn (which has not yet been directly associated with miR-1 regulation) was unchanged in our study, upregulation of these mRNAs was likely attributable to loss of miR-1 rather than an upregulation of miR-206. This interpretation is consistent with the observation that muscle development and regeneration were unaffected by knock-out of the miR-206/133b cluster, suggesting that they are expendable, perhaps due to functional overlap with the miR-1/133a clusters [72]. In the same context, muscle growth and adaptation were unaffected by manipulation of miR-206 levels in mice using adeno-associated viral vectors, suggesting miR-206 may not be a key regulator of these functions in post-natal muscle [73]. However, without further investigation, it is not possible to rule out a role for miR-206 to counterbalance miR-1 signaling in fine regulation of the muscle response after BoNT/A injection.
Contextualizing this miRNA expression pattern with the literature and downstream targets assessed in this study suggests a likely, but complex, role for muscle miRNAs in modulating muscle atrophy induced by transient muscle paralysis (Fig 6). For example, diminished miR-133a/b expression is consistent with upregulation of many factors involved in ECM remodeling in other tissues suggesting a potential role for these pathways in muscle structural changes. In turn, downregulation of miR-1 may be involved in the upregulation of miR-1/206 targets IGF-1 and Hdac4, and resultant signaling leading to NMJ degeneration and proteasomal degradation. Finally, we speculate that temporal upregulation of miR-206 may dynamically compete with miR-1 signaling for fine control of muscle atrophy responses.
The novel observation of upregulation of the differentiation inhibitor Hdac4 following BoNT/A-induced muscle paralysis is consistent with observations in other models of neuromuscular impairment. Enhanced Hdac4 expression has been reported after sciatic nerve transection, and in neurogenic muscle atrophy mouse models of ALS and NMD (Neuromuscular degeneration) [12], and upregulation of Hdac4 is associated with faster functional muscle decline and lower muscle re-innervation ability in ALS patients [74]. In addition, our evidence of elevated Myog expression, enhanced Atrogin-1 and MuRF-1 protein levels, and downregulated Dach2 and Fgfbp1 gene expression strengthen the likelihood that Hdac4 plays a significant role in BoNT/A-induced muscle atrophy, as they are consistent with earlier reports of denervation-induced changes [12][13][14]. Likewise, our observed decrease in expression of Hdac4 target gene Fgfbp1 after BoNT/A-induced muscle atrophy, is consistent with Hdac4-Fgfbp1 signaling after denervation [32]. Finally Hdac9, another target of Hdac4 [13], was not downregulated in our study, suggesting that while BoNT/A-and denervation-induced atrophy may share similar signaling pathways, they are distinct in some aspects.
Our observations should be weighed in the context of several general limitations. Given their known regulation in a variety of neuromuscular pathologies we focused on quantifying altered canonical muscle miRNA expression; however, other miRNA sequences found to be important in muscle diseases may also have a role in the response of muscle to transient paralysis [75] and our data provide a framework for genome wide analyses of the skeletal muscle response to BoNT/A. Also, we were not able to directly contrast miRNA responses following transient muscle paralysis to other models of muscle dysfunction because the timing, specificity, and extent of miRNA mediated responses are likely to vary across types of neuromuscular injury. Furthermore, our studies largely focused on gene expression changes rather than corresponding protein levels. Altered mRNA levels are not always proportional to protein responses, however, and the relative importance of changes in gene expression therefore cannot be gauged simply from the magnitude of fold change. Additionally, our study investigated short-term molecular responses after BoNT/A induced muscle paralysis without contemporaneous histological documentation of tissue changes. Studies directly relating miRNA/mRNA to histological evidence of muscle atrophy would be of potential relevance, given the widespread use of BoNT/A in the clinic. Finally, our study assessed mRNA and protein patterns in the context of miRNA alterations to survey candidate pathways for future mechanistic assessment, but did not functionally test regulation of these changes by the candidate miRNAs. This approach simplifies the complexity of miRNA regulation, in which feedback loops between signaling proteins such as Igf1 and Tgfb1 affect transcription of canonical muscle miRNAs [76,77] and miRNA can stimulate as well as inhibit target gene expression [78]. As such, we believe that future studies with genetic models will be required to establish a more direct link between miRNA changes and expression of Hdac4 and to determine the relative significance of other target genes that respond to intramuscular BoNT/A injection.
In summary, we have described molecular changes in the mouse quadriceps following BoNT/A-induced muscle paralysis. We observed, for the first time, differential regulation of muscle miRNAs from the miR-1/206 and miR-133 families and increased levels of Hdac4 and downstream mRNA targets following transient muscle paralysis of skeletal muscle. In the context of the literature, our BoNT/A-induced changes appear consistent with a potential role for miR-133/Tgf-β1/Ctgf in ECM remodeling and fibrosis previously observed in muscle after injection with this neurotoxin, as well as a role for a miR-1/Hdac4/Myog signaling axis in modulating muscle atrophy in this model of muscle paralysis.