FSHD Animal Model Progress

The FSH Society Patient Connect Conference was recently held at the Westin Copley Place Hotel in Boston. I am happy to report that real progress is being made in developing animal models that just may recapitulate the key elements of this, the most common adult muscular dystrophy. As a reminder, Fasioscapulohumeral muscular dystrophy (FSH, or FSHD) is caused by the aberrant expression of DUX4 in muscle. The gene for DUX4, a transcription factor, is located near the telomere of Chromosome 4. In FSHD Type 1, DUX4 expression in muscle is provoked due to deletion of adjacent D4Z4 repeated element units which comprise an epigenetic control architecture. In FSHD Type 2, the same DUX4 gene is transcriptionally activated again through derepression by virtue of mutations in SMCHD1, a gene on Chromosome 18 encoding a protein that enforces epigenetic repression in various locations including the D4Z4 repeat region. The emerging generation of animal models are all working from the notion that excess DUX4 protein is toxic (the leading pathogenic hypothesis) and so recapitulate that pattern with expression cassettes that are either induced in transgenic mice or virally delivered via AAV vectors. But here’s the rub: DUX4 and the D4Z4 locus is not found in mouse, the animal host of choice for most preclinical work (though other DUX family members are). So it is unclear if a protein restricted to humans (and only a few other species) and aberrantly induced in muscle via epigenetic derepression can recapitulate key elements of disease when induced in the mouse, a naive species. We should know a lot more about the performance of these surrogate disease models in the next year or so.

Earlier FSH mouse models also worked via the same concept of DUX4 expression but with two important differences: 1) high levels of DUX4 expression were driven using strong promoters and strong polyadenylation signal sequences such as that found in SV40. These expression systems worked yet they seemed to create a muscle pathology far worse than the FSH disease itself. 2) The expression cassettes used only the DUX4 coding region which eliminated potential RNA editing therapeutic targets that are found in the full length mRNA. In the new generation of models both of these problems are fixed by relying on expression cassette that include the full length DUX4 gene and employ the less efficient native polyadenylation signal sequence. As a result, DUX4 protein expression is reduced and the emerging pathology seems to look more like the humans disease (though perhaps a bit early to tell) and also to have a slower speed of progression that will permit easier testing of drug pharmacodynamics and functional outcomes. This is good news for any therapeutic concept that targets the DUX4 mRNA or protein directly.