Dual exon skipping in myostatin and dystrophin for Duchenne muscular dystrophy

Background Myostatin is a potent muscle growth inhibitor that belongs to the Transforming Growth Factor-β (TGF-β) family. Mutations leading to non functional myostatin have been associated with hypermuscularity in several organisms. By contrast, Duchenne muscular dystrophy (DMD) is characterized by a loss of muscle fibers and impaired regeneration. In this study, we aim to knockdown myostatin by means of exon skipping, a technique which has been successfully applied to reframe the genetic defect of dystrophin gene in DMD patients. Methods We targeted myostatin exon 2 using antisense oligonucleotides (AON) in healthy and DMD-derived myotubes cultures. We assessed the exon skipping level, transcriptional expression of myostatin and its target genes, and combined myostatin and several dystrophin AONs. These AONs were also applied in the mdx mice models via intramuscular injections. Results Myostatin AON induced exon 2 skipping in cell cultures and to a lower extent in the mdx mice. It was accompanied by decrease in myostatin mRNA and enhanced MYOG and MYF5 expression. Furthermore, combination of myostatin and dystrophin AONs induced simultaneous skipping of both genes. Conclusions We conclude that two AONs can be used to target two different genes, MSTN and DMD, in a straightforward manner. Targeting multiple ligands of TGF-beta family will be more promising as adjuvant therapies for DMD.


Background
Duchenne muscular dystrophy (DMD) is an X-linked recessive neuromuscular disorder, which is caused by dystrophin deficiency in muscle fibers. DMD fibers are more sensitive to muscle damage, leading to degeneration and replacement of muscle fibers by fat and connective tissue (fibrosis). Monaco et al found that frame shift mutations in the DMD gene will lead to a truncated and non-functional form of dystrophin [1], which become the primary cause of the disease. However, mutations which maintain the DMD open reading frame result in shorter dystrophin proteins that retain the essential actin binding-, cysteine rich-and carboxy terminal domains, and thus are partly functional [1].
Patients with such mutations develop the less severe Becker muscular dystrophy. This reading frame rule holds true for~91% of DMD cases [2] and has inspired the development of the exon skipping strategy, which employs antisense oligonucleotides (AONs). These small synthetic RNA molecules are complimentary to exonic or splice site sequences, thereby upon hybridization are able to modulate exon inclusion by the splicing machinery (recently reviewed in [3][4][5]).
Comprehensive studies done by our group and others have provided the proof of principle of the therapeutic feasibility of the AON to reframe dystrophin transcripts and restore dystrophin synthesis, both in vitro [6][7][8] and in vivo using the mdx and hDMD mice [9][10][11]. Subsequent clinical trials have shown that two different AON chemistries, either 2'-O-methyl phosphorothioate (2'OMePS) [12] or phosphorodiamidate morpholino oligomer (PMO) [13] targeting DMD exon 51 can restore local dystrophin synthesis in DMD patients with no to minimum side effect.
However, other major facets of DMD pathology which include severe muscle wasting, fibrosis and flawed muscle regeneration may reduce the efficacy of the DMD exon skipping therapy. In addition, as DMD patients suffer from muscle degeneration from their early life, myoblasts undergo extensive division in an attempt to regenerate, which eventually leads to exhaustion of the muscle regenerative potential [14][15][16]. Therefore, several additional therapies have been considered to overcome these problems, in which myostatin inhibition has received considerable interest.
A recent study by Dumonceaux et al [37] targeted the myostatin receptor AcvRIIb using shRNA and coupled it with a modified U7 small nuclear RNA to restore dystrophin in single adeno-associated vectors (AAV). This study showed that combination therapy has an added value in improving muscle physiology. However, as recently reviewed by Tang et al [38], one major limitation of AAV-based gene therapy approach is the immune response against the vector and/or transgene.
In this study we chose to use 2'OMePS AON to target MSTN exon 2 and disrupt its translational reading frame in order to knockdown its expression level. As it harbors the same chemistry modification as the PRO051 DMD AON currently in trial, we hypothesized that the administration of the two AONs as a cocktail would be achievable and may thus yield a combination treatment by simultaneously correcting the Dmd transcript and downregulating the Mstn transcript. In addition, as recently reviewed in [39], the 2'OMePS modification enhances the stability and increases the in vivo half-life compared to the previously studied butanol-tagged antisense [40] or siRNA-mediated RNA interference approaches targeting myostatin [41]. We evaluated its feasibility to induce exon skipping and downregulate MSTN expression in myotubes cultures and DMD mouse models. Furthermore, we combined it with several DMD AONs to look into the possibility of skipping two genes simultaneously ( Figure 1).

Antisense oligonucleotides
AON with phosphorothioate backbones and 2'-O-methyl ribose modifications used in this study were synthesized and high-pressure liquid chromatography purified by Eurogentec, Belgium. The sequences are listed in Table 1.

Animals and AON injections
All experiments were performed with 5-6 weeks old mdx mice under the approval of the Animal Experimental Committee (DEC07195) of the LUMC. In the single injection experiments, the AONs were administered into the gastrocnemius muscle at the dose of 40 μg per injection in 50 μl physiological salt. Control DMD AON (M23D) was injected in the contralateral muscle. After 4 days, the mice (n = 2) were sacrificed and muscles were isolated, snap frozen and sectioned at three different positions (proximal, medial and distal relative to the tendon) for total RNA isolation.
In the consecutive injections experiments, cocktails of MSTN or control AONs with M23 DMD AON were injected at the dose of 40+40 μg per injection into the gastrocnemius muscle. Four consecutive injections were performed with a 48 hours resting time between the second and third injection. The mice (n = 6) were sacrificed at 6 hours, 1 and 2 days after the last injection. The muscles were isolated, snap frozen and sectioned. Pool of sections from different area of the muscles was collected for RNA isolation.

RNA Isolation, RT-PCR and Quantitative PCR analysis
Cell lysates were prepared using lysis buffer provided in the NucleoSpin RNA II kit (Macherey-Nagel, Germany)  . Antisense oligonucleotides (AON) targeting exon 2 will hybridize and hide the exon from the splicing machinery, resulting in skipping exon 2 upon mRNA splicing. The removal of exon 2 will disrupt the open reading frame (ORF; *) and the protein will lack part of the propeptide and the entire mature domain (B). Dystrophin gene (DMD) consists of 79 exons, whereas the proteins consist of actin binding-, central rod, cysteine rich-and C-terminal domains (C). One of the examples of DMD is deletion in exon 48-50 which disrupted the reading frame and introduces premature stop codon (*). Due to this mutation, part of the central rod and the entire Cys-rich and C-terminal domains are missing (D). In the therapy (currently in clinical trials), AON is directed towards exon 51. With the similar principle, exon 51 will be skipped upon mRNA splicing which in turn restores the ORF. This internally-deleted DMD will be translated as dystrophin with shorter central rod domain. However, since the essential C-terminal domain is retained, the protein is partially functional (E). and RNA was isolated using the same kit according to manufacturer's instructions. Sectioned muscles were initially lysed using the same buffer and crushed with MagNA Lyser Green beads in the MagNA Lyser instrument (Roche Diagnostics, Germany) at 7000 oscillation speed for 2 × 20 seconds. RNA was isolated using the same procedure. The RNA quantity and integrity were measured using RNA 6000 Nanochip in the Agilent 2100 bioanalyzer (Agilent Technologies). cDNA was synthesized from 500 ng of RNA using RevertAid H Minus M-MuLV Reverse Transcriptase (MBI Fermentas) with 40 ng of random hexamer primer, according to the manufacturer's instructions. 10× diluted cDNA was then amplified by PCR using primers listed in Table 2

Myostatin AON induces exon 2 skip in vitro
Based on previously described guidelines [44], we designed three AONs targeting different regions in MSTN exon 2 gene and used previously designed AONs targeting DMD exons as controls ( Table 1). All AONs consisted of 2'-O-methyl RNA and had a phosphorothioate backbone in order to resist endonucleases and RNaseH degradation. We evaluated their efficiency using immortalized human myoblasts (7304-1) as well as primary myoblasts from a healthy individual (KM109) and a DMD patient with a deletion in exon 51-55 (DL589.2). We used differentiated myotube cultures because the expression levels of myostatin and dystrophin were found to be higher in the differentiation than in proliferation stage (data not shown).
To examine the feasibility to induce myostatin exon 2 skipping, different concentrations of each AON were transfected into the myotube cultures using the cationic polymer polyethylenimine (PEI). More than 80% of the cells showed specific nuclear uptake upon transfection with 5'-fluorescein (FAM)-labeled control AON ( Figure  2A). RT-PCR performed two days post transfection (Figure 2B) and subsequent sequencing analysis ( Figure 2C) showed the exclusion of exon 2 from the myostatin transcript in the myostatin AON-transfected cells, resulting in a premature stop codon formation. This internally truncated fragment was not observed in any of the non-transfected and control AON-transfected myotubes. One myostatin AON, namely AON1, gave the most consistent and highest skipping efficiency [Additional file 1]. Thus we further used the AON1 (addressed as myostatin AON from now on) and confirmed its exon skipping ability in human and to a lower extent in mouse cells models, using its perfect complementary to the human and mouse MSTN sequences ( Figure 2B and not shown).

AON-mediated exon skipping decreased myostatin transcript expression level and increased expression of myogenic regulatory factors
Next, we determined the expression level of MSTN transcript upon exon skipping by quantitative real time PCR. We used primers in exon 1 and exon 2, thereby detecting only the non-skipped products. As shown in Figure 2D, the levels of full length MSTN transcript decreased significantly in the myostatin AON-transfected samples, with slight variations in the knockdown levels between different cells. Notably, these decreases were achieved in an inverse dose-dependent manner, meaning that the most pronounced knockdown was achieved with 100 nM AON, which was the lowest concentration of myostatin AON tested. This effect was seen especially in the KM109 cells. In addition, the control AON induced less pronounced but still significant decreases of the full-length MSTN levels in some cells (7304-1 and DL589.2) ( Figure 2D), which might be due to the transfection. Therefore, the expression levels in myostatin AON-transfected samples were compared with those in the control AONtransfected samples. As outlined in Figure 1B, myostatin skipping results in premature stop codon in exon 3 and thus non-functional myostatin protein. Several studies have described that myogenic cells respond to myostatin by down-regulating the expression of key transcriptional regulators of muscle development such as Pax3, Pax7, p21, MyoD, Myf5 and Myog [23][24][25][26][27]45,46], explaining its inhibitory effects in differentiation. Therefore, inhibition of myostatin should lead to an increased expression of these myogenic regulators and thus can be used as a functional readout of the myostatin knockdown. Correspondingly, knockdown of myostatin in differentiated myotubes cultures resulted in a consistent upregulation of MYF5 and MYOG expression ( Figure 3A and 3B). The fold increase varied between 3-12 fold and dependent on the myoblast cultures used, which may reflect the different myogenic differentiation potential of the different myoblast cultures.

Dual exon skipping in MSTN and DMD in control and DMD patient derived myotubes
Our ultimate goal was to use AONs to simultaneously reframe the mutation in the DMD gene and to downregulate MSTN expression in order to correct the primary genetic defect as well as to enhance muscle regeneration. Therefore, we combined the myostatin AON with several DMD AONs used in the previous studies [6,7,47]. We specifically applied DMD AONs covering the hotspot mutated regions in DMD gene, namely h8AON3, h44AON1, h51AON1 and h54AON1 (Table  1) in control myotubes cultures (primary (KM109) or immortalized myoblasts (7304-1)). Myotubes were transfected with either myostatin or DMD AONs or a mix of both. As shown in Figure 4, MSTN exon skip was observed in all myostatin AON-transfected samples, regardless of the presence of DMD AON. Vice versa, all DMD AON-transfected myotubes showed the exon-specific dystrophin skips described before [6,7,47], regardless of the presence of MSTN AON. To investigate the possibility that the two AONs hybridized to each other rather than to their intended target, we varied the transfection conditions by mixing the AON before complexing with PEI, or preparing the AON-PEI complexes separately. However, no difference in their efficiency to induce skipping was found. We observed skipping with the total AON concentration of 200 nM (Figure 4 . RNA was isolated 2 days post-transfection. cDNA was synthesized using random hexamer (N6) primers and subjected for PCR using primers in exon 1 and 3 (B). Note the inverse dose-dependent skips in KM109 samples. Skip fragment was confirmed by sequencing analysis (C). Quantitative real-time PCR was performed using primers in MSTN exon 1 and 2, thereby depicting the expression of remaining full length or non-skipped transcript (D). Data are means ± SD from 3 to 4 independent experiments. Expression was normalized with GAPDH. Statistical analysis was performed using Student's t-test, using the 500 nM control AON-transfected samples as reference. *P < 0.05 the feasibility of simultaneous antisense-mediated skipping of exons of two different genes.
To further assess its therapeutic potential, we performed dual exon skipping in the DL589.2 cells, which were derived from a DMD patient with an exon 51-55 deletion. Our previous study has shown that the reading frame can be corrected by an exon 50 skip upon transfection with h50AON1 [7]. As shown in Figure 5

Myostatin exon skipping in vivo
We subsequently examined the ability of myostatin AON to induce exon skipping in the mdx mouse model. Myostatin AON was injected at a dose of 40 μg/injection into the gastrocnemius muscle of the mdx mice. DMD AON targeting exon 23 (M23D, previously denoted as M23D(+02-18) [9]) was injected into the contralateral muscle. This AON served as a positive control and was shown by many different groups including ours [9,11] to induce efficient and robust exon skipping. The mice were sacrificed at four days after injection. We classified different sections of the muscles as the proximal (P), medial (M) and distal (D) relative to the tendon and observed heterogenous patterns of Mstn exon 2 skips ( Figure 6A). The skipping levels in the proximal and medial parts of the muscles were higher than in the distal, while Dmd skips appeared to be heterogenous throughout the muscles. Our previous study in dystrophin exon skipping had indicated that higher skipping efficiencies could be obtained by multiple injections of the AON. In order to increase the efficiency of the myostatin AON and to answer whether dual skipping can be achieved in vivo, we injected cocktails of M23D DMD AON with myostatin or control AONs once a day for four times, with a 48 hours resting time between the second and third injections. RT-PCR analysis was performed from RNA isolated from proximal, medial and distal parts of the muscle ( Figure 6B). 6 hours after the last injection, we already observed weak myostatin skip in one of the two injected muscles. At the later time points, the skipping levels modestly increased, suggesting a trend of accumulations of skipping over time. Conversely, the Dmd skips showed relatively more stable and more importantly higher skipping levels.

Discussion
DMD patients suffer from progressive muscle wasting due to the absence of functional dystrophin protein.
The AON-mediated exon skipping therapy has been shown to successfully reframe the mutated DMD gene and restore local dystrophin synthesis in DMD patients [12,13]. This approach is currently viewed as one of the most promising therapeutic approaches for DMD. However, it targets dystrophin transcripts, which are produced only in muscle and not in adipose and fibrotic tissues, thus does not specifically aim to enhance muscle regeneration or reduce fibrosis levels. To address both the underlying genetic defect and the loss of muscle, we here explored the possibility of using AONs to inhibit myostatin expression and AONs to reframe dystrophin transcripts simultaneously.
Our results show that the second exon of the MSTN gene can be skipped in multiple myotube cultures, either derived from healthy individuals, DMD patient or mouse. The removal of exon 2 disrupts the open reading frame and introduces premature stop codon. We observed downregulation of myostatin transcript levels up to 50-80%, depending on the type of cells tested. One can expect the truncated transcripts to be degraded via nonsense mediated decay, thus not being translated into proteins. Nonetheless, we observed a clear skip product in the RT-PCR analysis, suggesting that the transcript might be fairly stable. However, the premature stop codon introduced will result in truncated myostatin protein that lacks~25% of the propeptide and the entire mature domains, which will abrogate its downstream signaling.
The effects of interventions with myostatin expression in muscle differentiation were assessed by increased expressions of MYF5 and MYOG. Although the expression levels of these genes were increased in all samples transfected with MSTN AON, there was no clear correlation between the fold decrease of MSTN and the fold increase of MYF5 and MYOG. One explanation for this observation could be that a threshold effect of myostatin knockdown on the expression of these genes is already attained when the MSTN transcript is slightly downregulated. However, the fact that we observed the upregulation of these genes in three to four independent experiments using different myoblasts suggests that this effect of myostatin downregulation using MSTN AON is valid and reproducible.
Finally, the mix of myostatin and dystrophin AONs showed that skipping two genes is feasible without interfering each other. The experimental set up was varied by mixing the AONs before complexing with the transfection reagent, or by preparing complex of AON and transfection reagent separately before mixing with each other to rule out the possibility that two AONs hybridize to each other. There seems to be no difference in the skipping efficiency, regardless if both AON were pre-mixed or not. However, we recommend to check the hybridization-likelihood by verifying the sequences of both AONs using freely available software tools such as RNAstructure [48].
One less encouraging finding was considerably low Mstn skipping level in vivo compared to the cultured cells. The M23D dystrophin AON, which was injected as a control in the contralateral muscle resulted in higher dystrophin skipping efficiency, confirming that the AON was injected properly. Repeated injections only showed modest improvements in the exon skipping level of myostatin. We also observed heterogenous skipping patterns throughout the muscles which seem to be specific for Mstn and not Dmd. We offer two reasons to explain the low myostatin skipping in vivo. First, we noted that the myostatin exon skipping efficiency in mouse cells was lower than in human cells in vitro. Thus, there is a possibility of species-dependent differences in the efficiency of our myostatin AON, although the AON was designed for both human and mouse. Another explanation might be that myostatin is also expressed in muscle satellite cells and fibroblasts. In vitro, we indeed saw higher MSTN expression in primary cultures containing mix populations of fibroblasts and myoblasts than in immortalized myoblasts. We previously demonstrated that the leaky muscle membrane facilitates the AON to enter the muscle fibers, where dystrophin is predominantly expressed, thus explaining the high efficiency of dystrophin AON. The delivery of the AON to fibrotic area and satellite cells, where myostatin is also expressed will presumably be more difficult. Furthermore, a recent study by Kang et al [49] which also targeted myostatin exon 2 showed efficient exon skipping in vivo using octa guanidine morpholino oligomers. The chosen chemistry seems to be the key to facilitate the delivery and thereby inducing prominent exon skipping in vivo. Finally, several studies have shown that members of the TGF-beta family other than myostatin are responsible in inhibiting muscle differentiation [35,50,51]. Therefore, as a future prospective, targeting multiple ligands seems to be a more promising strategy, especially when combined with dystrophin restoration.

Conclusions
In conclusion, our results provide a rationale for the use of two AONs targeting two different genes and further extend the versatility of exon skipping-based therapy. The efficiency and delivery techniques remain to be improved, especially for further in vivo studies. While the correction of genetic defect is essential, combination with other adjuvant therapies which do not rely solely on myostatin inhibition, but also other TGF-beta family members regulating muscle differentiations might be more beneficial for DMD therapies.

Additional material
Additional file 1: Human primary control (KM109) myoblasts were differentiated for 7 days before transfection with 3 different MSTN AONs at 500 and 100 nM concentrations. All AONs were designed to target exon 2 of the myostatin gene. The sequences are listed in table 1. RNA was isolated 2 days post-transfection. cDNA was synthesized using