Antisense PMO cocktails effectively skip dystrophin exons 45-55 in myotubes transdifferentiated from DMD patient fibroblasts

Antisense-mediated exon skipping has made significant progress as a therapeutic platform in recent years, especially in the case of Duchenne muscular dystrophy (DMD). Despite FDA approval of eteplirsen–the first-ever antisense drug clinically marketed for DMD–exon skipping therapy still faces the significant hurdles of limited applicability and unknown truncated protein function. In-frame exon skipping of dystrophin exons 45–55 represents a significant approach to treating DMD, as a large proportion of patients harbor mutations within this “hotspot” region. Additionally, patients harboring dystrophin exons 45–55 deletion mutations are reported to have exceptionally mild to asymptomatic phenotypes. Here, we demonstrate that a cocktail of phosphorodiamidate morpholino oligomers can effectively skip dystrophin exons 45–55 in vitro in myotubes transdifferentiated from DMD patient fibroblast cells. This is the first report of substantive exons 45–55 skipping in DMD patient cells. These findings help validate the use of transdifferentiated patient fibroblast cells as a suitable cell model for dystrophin exon skipping assays and further emphasize the feasibility of dystrophin exons 45–55 skipping in patients.


Introduction
Duchenne muscular dystrophy (DMD) is a lethal, progressive myopathy affecting approximately 1 in every 3600-5000 male births and is caused by deleterious mutations in the dystrophin (DMD) gene [1][2][3][4]. Mutations in DMD can also cause another milder form of muscular dystrophy known as Becker muscular dystrophy (BMD) [5]. Typically, DMD arises from outof-frame mutations (~65% of patients) while BMD generally arises from in-frame mutations (~82% of patients) [6][7][8]. PLOS  The observation that truncated dystrophin protein arising from in-frame mutant DMD transcripts could still maintain partial functionality-as in the case of BMD-helped provide the rationale for a therapeutic approach involving splice modulation via synthetic polymers. Antisense oligonucleotides (AOs) are chemically modified nucleic acids which can hybridize to pre-mRNA and can affect splicing and protein synthesis [9,10]. By utilizing AOs, mutationcarrying exons and flanking exons can be selectively removed or "skipped" from the final messenger transcript, restoring the reading frame and producing a truncated protein which may retain some functionality-in essence, exon skipping could convert a DMD phenotype to a BMD phenotype [9,11,12]. Several in vitro and in vivo studies have now demonstrated the feasibility of antisense-mediated exon skipping in DMD [13][14][15][16][17][18] and the first-ever clinical AO drug for treating DMD, eteplirsen (Exondys 51), was approved by the FDA in 2016 [19].
Exon skipping as a therapy for treating DMD is not without its challenges. The first major drawback is the requirement for specifically-targeted AO sequences for any given exon. This means a great deal of time and money needs to be spent in evaluating individual AO sequences to address a wide range of patient mutation patterns. Another big challenge is determining the stability and functionality of the truncated proteins [20,21]. A solution to the specific issue of validating oligo sequences for accommodating multiple mutation patterns is using analytical software algorithms to predict AO exon skipping efficiency, as has been described [22]. In conjunction with in silico predictive tools, another possible avenue to circumventing these issues is multi-exon skipping of DMD exons 45-55. First, in terms of applicability, a large proportion of all DMD patients (~47%) harbor mutations within this "hot-spot" region of exons 45-55, and up to~63% of DMD patients with deletion mutations could benefit from skipping exons 45-55 [23,24]. Multi-exon skipping of exons 45-55 would also address the issue of unknown truncated protein stability/function, as patients exhibiting this particular pattern of deletion mutation present with exceptionally mild symptoms or are asymptomatic [23][24][25].
As more AOs are required to bind to the same transcript to ensure maintenance of the open reading frame, dosage levels must be sufficiently high enough to facilitate skipping multiple exons while avoiding overt toxicity. While several AO chemistries have been designed and tested [9,26], one of the most promising antisense chemistries developed is the phosphorodiamidate morpholino oligomer (morpholino or PMO). The stability, safety, and tolerability of PMOs have been well-documented [27][28][29], even at high doses [30,31], and it is worth noting that the clinically-utilized eteplirsen is a PMO chemistry [32]. PMOs could therefore address issues regarding the potential toxicity of AO cocktails as they could be delivered at concentrations high enough to facilitate exons 45-55 skipping while avoiding toxic effects. Our group has demonstrated this previously in a dog model of DMD, using a cocktail of PMOs to effectively skip multiple exons in vivo without overt toxicity [14].
The first step toward clinical utility of any AO drug is the demonstration of its efficacy in vitro; thus, the establishment of a suitable in vitro exon skipping assay for determining dystrophin exon skipping is essential. To this end, muscle cell types (myoblasts and myotubes) expressing dystrophin are typically used. However, the collection of patient muscle tissue via biopsy is highly invasive, requires specialized equipment and preservation techniques [33], and yields only a small amount of what are often poorly proliferative cells [34]. Myotubes converted from fibroblasts via MYOD1 transduction express dystrophin at levels sufficient to determine the effectiveness of AOs at facilitating DMD exon skipping and protein rescue [13] and represent an alternative approach to biopsied patient muscle tissue at evaluating exon skipping efficiency in vitro.
In this study, we tested the exons 45-55 skipping efficacy of PMO sequences designed using a predictive software tool by transfecting PMO cocktails into transdifferentiated DMD patient myotubes harboring deletion mutations amenable to reading frame correction via exons 45-55 skipping. We observed a dose-dependent production of exons 45-55 skipped transcripts as well as the rescue of dystrophin protein in DMD patient cells. This is the first-ever demonstration of robust dystrophin exons 45-55 skipping in transdifferentiated patient cells.

MYOD1 transduction and cell sorting
A pRetroX-IRES-ZsGreen1 expression vector (Clontech) containing the human MYOD1 coding sequence, a pVSV-G envelope vector, and a gap-pol expression vector were prepared as described previously [13]. These vectors were co-transfected into HEK 293T cells via standard calcium phosphate transfection. Viral supernatant was collected after 48-72 h incubation and added to 70-80% confluent DMD fibroblast cells (40 mL viral supernatant in a T225 flask) along with 8 μg/ml polybrene (Sigma-Aldrich). DMD fibroblasts were incubated for 24 h at 37˚C, then the viral supernatant was replaced with fresh growth media and cells were incubated for an additional 48 h. Cell sorting via FACS was performed on a FACS Area III flow cytometry system (BD Bioscience) by the Faculty of Medicine and Dentistry's Flow Cytometry Core at the University of Alberta. Sorted ZsGreen-positive cells were seeded at 1x10 5 cells in 500 μl of total volume into each well of a 12-well collagen-coated plate and incubated in growth media for 24 h. Culture media was then changed to differentiation media and cells were incubated until myotubes formed, with differentiation media being replaced every 2-3 d.

Design and transfection of PMOs
The in silico design of 30-mer AOs targeting DMD exons 45-55 was performed using a predictive software algorithm developed by our group [22] and PMOs were synthesized by Gene Tools (Oregon, USA). PMOs at 1, 3 or 10 μM each were transfected as a cocktail into differentiated DMD myotubes using 6 μM Endo-Porter transfection reagent (Gene Tools). Cells were incubated with PMO for 48 h and then media was changed to fresh media and cells were incubated an additional 72 h before harvesting for analysis.

RT-PCR analysis and sequencing
Total RNA was collected from cells using Trizol (Invitrogen) and 200 ng of total RNA was used for analyzing exons 45-55 skipping efficiency via SuperScript III One-Step RT-PCR System with Platinum Taq DNA Polymerase (Invitrogen). Primers for dystrophin: FWD GACAAGGGCGATTTGACAG; REV TCCGAAGTTCACTCCACTTG. Primers for GAPDH: FWD TCCCTGAGCTGAACGGGAAG; REV GGAGGAGTGGGTGTCGCTGT. PCR products were electrophoresed on a 1.5% agarose gel and bands were excised using a Wizard1 SV Gel and PCR Clean-Up kit (Promega). Sanger sequencing of excised bands was performed by The Applied Genomics Core at the University of Alberta.

Patient mutation analysis and exon skipping approach
DMD fibroblast cell lines GM05162 and GM05017 supplied by the Coriell Institute were originally sampled from two clinically affected males aged 13 and 12, respectively. Both individuals exhibited progressive muscle weakness and were wheelchair bound before or at age 10. GM05162 harbors a deletion of dystrophin exons 46-50, resulting in an out-of-frame product which requires a cocktail of 6 PMOs and skipping of exons 45, 51-55 to restore the reading frame (Fig 1A). GM05017 contains an out-of-frame deletion of exons 45-50, requiring a cocktail of 5 PMOs and skipping of exons 51-55 (Fig 1B). Based on these mutation patterns, we utilized our exon skipping predictive tool [22] to calculate the expected exon skipping efficiencies for PMO sequences of 30-mer length covering all possible target sites for the corresponding exons. PMO sequences used are described in Table 1. For our oligo design, we also considered the binding free energy values between AOs and selected oligos with >-9 ΔG ( Table 2). The inframe skipping of dystrophin exons 45-55 produces a truncated product containing a hybrid rod repeat that is known to retain similar function to the full-length protein (Fig 1C).

MYOD1 transduction of DMD fibroblasts and conversion to myotubes
An expression vector containing ZsGreen and the human MYOD1 coding sequence was delivered via retrovirus into human DMD patient fibroblast cells and healthy human fibroblasts (Fig  2A) [13,35]. Following transduction, ZsGreen-positive cells were sorted via flow cytometry (Fig 2B) and seeded into collagen-coated 12-well plates. After adherence, cells were cultured in reduced-serum media to induce myogenic differentiation. Morphological examination showed that ZsGreen-positive MYOD1-transduced cells had become elongated and contained multiple nuclei (Fig 2C)-hallmarks of myotube morphology. To confirm successful transdifferentiation of fibroblasts to muscle cell type, we performed immunostaining for several markers of muscle identity. These cells expressed muscle-specific proteins myosin heavy chain and desmin, and healthy cells expressed dystrophin (Fig 2D). We then utilized a time-course expression assay to compare dystrophin mRNA expression between transdifferentiated healthy and patient cells. In both healthy and DMD patient cells, dystrophin mRNA expression was detectable by RT-PCR as early as 3 d after addition of differentiation media (Fig 2E).

Antisense-mediated multi-exon skipping of dystrophin exons 45-55 in DMD patient cell lines using PMO cocktails
Based on respective mutation patterns, skipping of dystrophin exons 45-55 in cell line GM05162 requires a cocktail of 6 PMOs, while skipping of exons 45-55 in cell line GM05017 requires a cocktail of 5 PMOs (Fig 1A and 1B). Following transfection of PMO cocktails, RT-PCR analysis showed exon skipped products of the expected molecular weight in both cell lines in a dose-dependent manner (Fig 3A). Sanger sequencing of exon-skipped products revealed that the skipped products contained in-frame concatenations of DMD exons 44 and 56 (Fig 3A). Immunostaining showed some rescue of dystrophin protein in both PMO-treated cell lines (Fig 3B).

Discussion
In this study, we demonstrated the feasibility of skipping dystrophin exons 45-55 in vitro using human DMD patients' myotubes converted from fibroblasts. This is the first successful demonstration of robust, dose-dependent dystrophin exons 45-55 skipping in DMD patient cells. While earlier attempts at exons 45-55 skipping in patient cells were unsuccessful [36], our results emphasize the importance of AO sequence optimization and highlight the utility of in silico predictive screening for potential AO sequences.  This is also the first demonstration of exons 45-55 skipping in transdifferentiated cells, which underscores the suitability of using patient fibroblast cells as an alternative to cells obtained via muscle biopsy for evaluating exon skipping. This method of measuring exon skipping efficiency in transdifferentiated myotubes offers advantages to other assays, such as easy monitoring of MYOD1 transduction efficiency via ZsGreen signal and effective induction of dystrophin expression, which can be difficult to induce [37].  Our group previously reported the efficacy of antisense-mediated dystrophin exons 45-55 skipping and rescue of dystrophin protein in vivo using a cocktail of vivo-morpholinos (vPMOs) in a mouse model of DMD [14,15]. Before such new-generation antisense oligos can be effective and safe in vivo they require rigorous sequence optimization in vitro. Here, we emphasize an effective method for in vitro assessment of exon skipping efficacy in transdifferentiated human DMD patient cells which can pave the way for subsequent in vivo and clinical studies. By utilizing muscle cells obtained through fibroblast transdifferentiation, researchers can access an effective cell model for assessing the exon skipping ability of novel antisense chemistries and gene sequence targets while avoiding challenges associated with utilizing muscle harvested from patient biopsies, such as the limited availability and poor proliferative ability of patient muscle samples [34,38].
Currently, the only clinically available exon skipping therapy is Sarepta's exon 51 skipping drug, eteplirsen (Exondys 51). Notwithstanding FDA approval of the drug in 2016 [19], eteplirsen has remained surrounded by controversy, with concerns being raised as to its clinical efficacy [39]. Furthermore, exon 51 skipping is limited in its therapeutic applicability, with an estimated~13% of all DMD patients being able to benefit from such an approach [40]. Several antisense-mediated exon skipping approaches are currently being evaluated across various clinical trials, targeting DMD exons 44 (NCT02958202), 45 (NCT02667483), 51 (NCT03375255), and 53 (NCT03167255); the respective therapeutic applicability of these targets is~6%,~8%,~13%, and~8% [40]. Notably, while the combined applicability of the aforementioned exon skipping approaches is~35%, a single exon skipping approach targeting exons 45-55 would theoretically be amenable to~47% of all DMD patients [24]. In addition to increased patient applicability, another advantage of skipping DMD exons 45-55 is that the resulting truncated protein is remarkably stable, as evidenced by patients harboring exons 45-55 deletion mutations who are either asymptomatic or display exceptionally mild symptoms [24].
In conclusion, our findings support the feasibility of future translation of the dystrophin exons 45-55 skipping approach into clinical practice for treating DMD. Potential hurdles such as possible side-effects of intermediate transcripts, toxicity assessment, and optimized PMO cocktail delivery will need to be addressed in future investigations. The development of a drug cocktail approach may also necessitate modification of existing regulatory body guidelines which currently consider individual AO sequences to be separate drugs. According to existing regulations, each individual component of a cocktail drug approach, including all possible combinations within that cocktail, would be required to undergo rigorous toxicological testing [41]. This creates significant barriers to the development of a drug cocktail approach which are both expensive and time-consuming to surmount.