FSHD is caused by inappropriate expression of the double homeobox (DUX) gene, DUX4. Because DUX4 protein is not generally detected in biopsy specimens from FSHD patients, it has been proposed that DUX4 initiates a disease process, including potentially persistent expression of target genes, that persists after a burst of DUX4 expression. One of the strongest DUX4 target genes is DUXA, itself a double homeodomain protein. Because DUXA is detected in the absence of DUX4 in FSHD patient cells, it has been suggested that DUXA may mediate DUX4 effects after DUX4 is no longer present in FSHD. DUX proteins are unique to eutherian mammals, transiently expressed in early blastomere-stage embryos, and comprise 3 clades (DUXA, B, C), of which DUX4 belongs to the DUXC clade.
Here, we investigate the 3 human DUX family representatives, DUXA, DUXB, and DUX4, as well a single homeodomain protein, sDUX, from platypus, a near non-eutherian relative, lacking DUX genes.
We find that DUXA and DUXB lack transcriptional activity. However a DUXA-VP64 synthetic transactivator produced transcriptome and chromatin accessibility profiles that overlap those of DUX4, including expression of ZGA genes and LTR elements. The DUXA-VP64 fusion is cytotoxic to myogenic cells and inhibits myogenic differentiation, while native DUXA (without fusion to VP64) is not toxic, and in fact counteracts the transcriptional activity of DUX4 on its target genes, including in FSHD patient cells. Since DUXA is one of the strongest DUX4 target genes, this activity potentiates feedback inhibition, shaping and limiting the activity of DUX4, opposite to what has been proposed. We further demonstrate derived and innovated functions of the DUX proteins vis-à-vis the putative evolutionary progenitor, sDUX.
The DUX gene family therefore comprises cross-regulating members of opposing function, with implications for their roles in FSHD.