Myotonic dystrophy (DM) is the most common muscular dystrophy in adults, affecting approximately 1 in 8000 individuals worldwide, with no currently available disease-modifying therapies. The primary pathomechanism in DM is a toxic RNA gain-of-function caused by non-coding microsatellite repeat expansions in the 3’UTR of DMPK (type 1 and congenital) or intron 1 of CNBP (type 2). The mutation leads to accumulation of repeat-expanded (CUG)n or (CCUG)n RNA which binds to and causes functional sequestration of muscleblind-like (MBNL) proteins. MBNL proteins are important for pre-mRNA processing, including alternative splicing, cellular localization, and polyadenylation, and their sequestration leads to altered expression and mis-splicing of hundreds of genes. Multiple MBNL-dependent mis-splicing events have been directly linked to specific DM phenotypes and have a quantifiable relationship to disease severity (1).
Here we co-opted well-characterized splicing events to generate DMXon, a disease-responsive platform to control the production of therapeutic protein. Under normal MBNL activity, as in non-disease affected tissues, production of the therapeutic product is inhibited by exclusion of an appropriate translation initiation sequence. As functional MBNL levels decline, cassette exon inclusion introduces this sequence to enable protein production. DMXon transgenes derived from different splicing events show dose-dependent responses to MBNL levels with variable dynamic ranges, therapeutic output, and self-regulation in vitro and in vivo.
By responding to disease state, DMXon transgenes have the potential to deliver individual, graded responses encountered in patients and allow for adaptation to disease progression and avoid toxicities associated with transgene over-expression when delivered using long-term therapeutic modalities, such as adeno-associated virus (AAV).