Set-up of an in vivo model of facioscapulohumeral muscular dystrophy (FSHD) based on human perivascular cells

DUX4, the best candidate gene of FSHD, is specific of human cells with mouse models expressing variant forms of DUX family. Moreover, human and mouse muscle express a different pattern of hormone receptor. To study the muscle regenerative potential of FSHD individuals and the effect of hormones as well as other substances in this process, we set up an in vivo model based on human muscle precursor cells, perivascular cells or pericytes (PVCs). These cells when derived from intra-muscle blood vessel have the ability to differentiate in myotube. We show that DUX4 is active in these cells and is able to induce its target genes and to impair muscle differentiation. Moreover, PVCs express an estrogen receptor pattern similar to that of human muscle cells. Accordingly, estrogens are able to reduce DUX4 transcriptional activity. Of importance, this reduction occurs only in differentiating PVCs whereas 17-? estradiol is ineffective towards DUX4 transcriptional activity during cell proliferation, thus recapitulating previous results obtained with myoblasts from FSHD patients. To evaluate PVCs survival, growth and differentiation, we analysed implants of PVCs in different districts of NSG mice: subcutaneous back, subcutaneous muscle and intramuscular. The results demonstrate that PVCs growth is similar in the different tissue districts whereas their muscle differentiation is increased when injected into injured muscle. Moreover, two different hydrogels were compared for in vivo PVCs implantation: a hydrogel consisting of photopolymerizable polyethylene glycol fibrinogen (PEG-FB) and the Matrigel a a gelatinous protein mixture derived from Engelbreth-Holm-Swarm (EHS) mouse sarcoma. The PEG-FB resulted less able to sustain PVCs growth and differentiation. Finally, we tested different markers to follow PVCs growth/survival and DUX4 activity. Cherry signal resulted the more stable over time whereas luciferase signal under the promoter of ZSCAN gene resulted more effective in monitoring DUX4 activity when compared to RFLP4b gene. Overall, these experiments have allowed to set up a model that will be used to monitor in vivo DUX4 function in muscle regeneration and to test the effect of different factors on it.