Several groups in academia and industry are pursuing nucleic acid-based strategies to inhibit DUX4 gene expression at the RNA level. Although there were some differences among the various strategies, particularly in the mechanism of action and delivery strategy, each shared a common feature: the requirement for the therapeutic nucleic acid species to form antisense base pairs with the DUX4 mRNA. The speakers in this session described 3 different mechanisms to accomplish DUX4 mRNA silencing with antisense approaches:

1.

RNA interference (RNAi) strategies utilizing small interfering RNAs (siRNAs) designed to bind DUX4 mRNA and trigger its degradation by the RNA-induced silencing complex (RISC) (Barbora Malecova, Avidity Biosciences; Katelyn Daman, University of Massachusetts; Jonathan Van Dyke, Arrowhead; Lindsay Wallace, Nationwide Children’s Hospital)

2.

Induction of RNAse H cleavage via RNA/DNA duplex using DNA gapmer antisense oligonucleotides (gapmer ASOs, or gapmers) (Linde Bouwman from Leiden University Medical Center)

3.

Steric hindrance using phosphorodiamidate morpholino oligomers (PMOs or Morpholinos) (Ngoc Lu-Nguyen, Royal Holloway University, and Nelson Hsia, Dyne)

Historically, the delivery of synthetic nucleic acids to the muscle has been challenging [16]. The presenters in this session described various methods to increase nucleic acid delivery to the muscle, including formulating nucleic acids with lipids (Linde Bouwman, Leiden University Medical Center; Katelyn Daman, University of Massachusetts), cell-penetrating dendrimers (Ngoc Lu-Nguyen, Royal Holloway University), or linking with antibodies or ligands targeting the Transferrin receptor (TfR) or other receptors present on muscle membranes (Barbora Malecova, Avidity Biosciences; Nelson Hsia, Dyne; Jonathan Van Dyke, Arrowhead). Like other traditional small molecule therapies, those employing ASOs and siRNAs require repeated lifelong administration to maintain an intended therapeutic effect. Alternatively, AAV-based muscle gene therapy systems use viral capsids to achieve widespread muscle delivery and can be designed to enable long-term expression following one administration (Lindsay Wallace, Nationwide Children’s Hospital). Current limitations of AAV vectors for muscle gene therapy include high production costs, the inability to adjust dose and readminister once a vector is delivered, and safety concerns associated with systemic delivery of high vector doses.

Linde Bouwman (Leiden University Medical Center) reported results from a study using uninduced ACTA1-MCM;FLExDUX4 mice, where DNA gapmers, mixed with Palmitate (C16), were delivered twice a week at 50 mg/kg for 4 weeks, followed by 50 mg/kg weekly for another 5 weeks. Compared to ACTA1-MCM;FLExDUX4 mice treated with a control gapmer, DUX4-gapmer-treated animals showed reductions in DUX4 and 3 DUX4-activated mouse target genes, and evidence of histological and functional improvement (reduction in myofibers with central nuclei from 26 to 18%, and 30% increase in 1250 meter treadmill performance, respectively). However, no improvements were found in muscle weight or strength measured using a hanging grid test and a grip strength test. This study is now published [17].

Barbora Malecova (Avidity Biosciences) presented in vitro and in vivo data using DUX4 target genes as a surrogate indicator of silencing efficacy, following treatment with DUX4-targeting siRNA conjugated to a murine Transferrin receptor (TfR) monoclonal antibody. Specifically, Dr. Malecova reported reductions in 4 DUX4 target genes (QPCR) and SLC34A2 protein (immunofluorescence) in 11 different human FSHD myotube lines treated with 10 nM siRNA. Similarly, dose-dependent reductions in 4 DUX4-activated mouse target genes were found in FLExDUX4 mice treated with siRNA+TfR antibody conjugate, 3 weeks later. No functional or histological outcomes were reported for the mouse study (Table 1).

Table 1 Silencing mechanism, nucleic acid species, and muscle delivery systems described in session 6 of the FSHD IRC meeting. aRef 19. bRef 20. cRef 18. dUsed FM10, a previously published sequence targeting the DUX4 polyA signal. Although the PMO operates via steric hindrance, blocking polyadenylation could lead to DUX4 mRNA instability and degradation (Ref 18)

Katelyn Daman (University of Massachusetts) used DUX4-targeting siRNAs formulated with docosanoic acid to facilitate delivery to FSHD myoblasts, FSHD myotubes, and an FSHD xenograft model. In vitro treatment of FSHD cells with 3 different concentrations of siRNA (0.5, 1, and 2 μM) produced dose-dependent reductions in 3 human DUX4 target transcripts. Similarly, 4 DUX4 target genes were reduced in FSHD xenograft mice treated subcutaneously with two 20 mg/kg doses of siRNA/docosanoic acid over a week.

Nelson Hsia (Dyne Therapeutics) described a strategy to deliver the DUX4-targeting PMO FM10 to FSHD myotubes by coupling the therapeutic oligonucleotide to a TfR antibody Fab fragment. FM10 had been previously reported to bind atop the DUX4 polyA signal, thereby potentially operating to destabilize the DUX4 mRNA by blocking polyadenylation [18]. Dr. Hsia reported reductions in 3 DUX4 target genes in human FSHD myotubes treated with 8 nM of the FM10-TfR antibody Fab fragment conjugate.

Ngoc Lu-Nguyen (Royal Holloway University) presented her work describing an in vitro screen of several PMOs targeting DUX4, followed by coupling of a lead sequence to a guanidinium dendrimer for in vivo delivery (Vivo-PMO). In vitro, several PMOs caused reductions in DUX4 and 3 target genes in FSHD myotubes. In vivo studies were performed in tamoxifen-induced ACTA1-MCM;FLExDUX4 mice, where animals received weekly intraperitoneal doses (10 mg/kg) of the lead Vivo-PMO for 30 days. This treatment led to partial reductions in DUX4 and 2 DUX4-activated mouse target genes and partial improvement in histological and functional outcomes. This study is now published [19].

Jonathan Van Dyke (Arrowhead Pharmaceuticals) presented a comprehensive study of DUX4 inhibition involving the delivery of a nucleic acid with an unspecified delivery system. The data reported suggested that the siRNAs and delivery system were highly effective at silencing DUX4 and improving several phenotypes. Specifically, in vitro delivery of 1, 10, and 100 nM DUX4-targeting siRNAs to human FSHD myotubes caused dose-responsive reductions in 10 DUX4 target genes. For in vivo studies, Arrowhead used the tamoxifen-induced ACTA1-MCM;FLExDUX4 model. Animals were treated with conjugated siRNA on days 3 and 5 after tamoxifen induction, followed by weekly injections thereafter for up to 30 days. This treatment regimen caused reductions in DUX4 and a DUX4-activated mouse target gene, improvement in body weight at 30 days, and increased performance on the rotarod 22 days after DUX4 induction. Improvements in fibrosis were also reported but not quantitated.

Lindsay Wallace (Nationwide Children’s Hospital) reported her progress on translating an AAV-based gene therapy project aimed at inhibiting DUX4 with an engineered microRNA called mi405 [20, 21]. Dr. Wallace reported several in vivo studies in both the uninduced and tamoxifen-induced TIC-DUX4 mouse models, which recapitulate mild and more severe forms of DUX4-related myopathy, respectively [22]. To address the durability of treatment, a single intramuscular dose of 1 × 1011 vector genomes (vg) of AAV6 serotype vectors carrying a U6.mi405 expression cassette protected muscles from histological damage out to 1 year (25% central nuclei in untreated limbs; 5% in treated limbs), and systemic intravenous (IV) delivery of 3 × 1013 or 3 × 1014 vg/kg improved activity, hindlimb rearing and rotarod performance out to 6 months. Similarly, tamoxifen-induced TIC-DUX4 mice treated IV with AAV6 or AAV9 serotyped U6.mi405 vectors showed no decline in cage activity over a 10-week period. Finally, in a collaborative study with the Emerson lab (University of Massachusetts) using an FSHD xenograft model, IM injection of AAV6 vectors reduced DUX4 and 8 DUX4 target genes.

Finally, Dr. Yi-Wen Chen, who presented an ASO study at the 2020 IRC meeting that is now published, did not present an abstract this year but participated in the Q&A at the end of this session [23].