Barking up the right tree: Human DUX4 and dog DUXC similarities highlight potential muscular dystrophy model

During embryonic development, waves of gene expression help pattern and grow the embryo. Many of these genes serve important roles throughout development and also later to maintain our tissues. Other genes are only active in the initial phases of embryogenesis and if activated inappropriately later in life, can wreak havoc on our bodies. Such is the case with DUX4, a double-homeodomain transcription factor encoded by a retrogene, a piece of DNA copied back from the mRNA of a parental gene by reverse transcription. Retrogenes have been known to drive evolution by insertion into genomic regions that alter their expression pattern. Interestingly, a recent study in Human Molecular Genetics from the Tapscott Lab, part of the Human Biology Division at Fred Hutch, shows that another consequence of retrotransposition might be to restrict gene expression to a single isoform. In normal development, DUX4 is expressed in the germline and in the early embryo where it drives the first wave of gene transcription but then must be silenced in later embryonic stages and adult tissues. The Tapscott Lab has found that if DUX4 becomes inappropriately expressed in muscle cells, it causes muscles to gradually weaken over time leading to fascioscapulohumeral muscular dystrophy (FSHD). It would seem that if expression of the DUX4 retrogene had such devastating consequences, then there might be mechanisms to regulate its activity at different time windows or in different tissues during our development. Dr. Stephen Tapscott describes that trying to understand this mystery of how DUX4, “has evolved to drive conserved developmental programs and maintain its expression, […] while minimizing deleterious effects”, is part of what drives him and his lab’s research.

Human DUX4 and mouse Dux retrogenes are derived from the parental gene DUXC. While these retrotransposition events in human and mice are thought to have occurred separately, both primates and rodents have lost the intron-containing, parental DUXC gene. The DUX4 and Dux retrogenes have similar but distinct binding sites suggesting a rapid divergence following retrotransposition. Interestingly, dogs, pigs and other Laurasiatherians have retained the parental DUXC gene, but its functional relation to the human DUX4 or mouse Dux retrogenes was unknown. In the Tapscott Lab’s recent study led by Dr. Chao-Jen Wong, the authors investigate the functional relationship between human DUX4, mouse Dux and canine DUXC. Since human DUX4 and mouse Dux are both expressed in the testes, they first looked for DUXC transcripts in canine testes to determine if multiple DUXC isoforms were expressed. In addition to the predicted double-homeodomain DUXC transcript, they identified an even more abundant isoform with insertion of an alternative exon (DUXC-ALT) that disrupts the first homeodomain. Comparing their structural domains, the Tapscott group found high similarity among the first homeodomain of human DUX4 and canine DUXC, with a 75% overlap in amino acid sequence, but found a lesser degree of overlap between DUX4 and canine DUXC-ALT. The authors then generated stable canine myoblast cell lines expressing the four transcription factors- human DUX4, mouse Dux, canine DUXC and DUXC-ALT- to directly compare their binding sites and transcriptional programs. They identified the genomic regions bound by the different DUXC-related proteins and, surprisingly, found that canine DUXC bound to sites that are a composite of the human DUX4 and the mouse Dux site, whereas DUXC-ALT favored binding to sites more similar to mouse Dux. The researchers then investigated the gene expression signatures regulated by canine DUXC and found that it is capable of driving early pluripotent transcriptional programs similar to both human DUX4 and mouse Dux, with the highest degree of correlation being between canine DUXC and human DUX4. Remarkably, they also uncovered that DUXC induces expression of FSHD signature genes and retrotransposons, again observing more overlap with canine DUXC and human DUX4 induced genes than compared to mouse Dux. Interestingly, DUXC-ALT was unable to initiate robust transcriptional responses. The striking similarities between human DUX4 and canine DUXC, in their structural domains, binding sites and transcriptional responses, including activation of FSHD signature genes, support the possibility that DUXC-containing mammals may make good preclinical models of FSHD.

DUXC and the retrotransposed human DUX4 and mouse Dux
Image provided by Dr. Chao-Jen Wong

While the Tapscott Lab investigated DUXC in canine cells, they are not proposing to use dogs as models for FSHD. Rather, their research shows compelling evidence that more generally, the use of mammalian animal models that contain DUXC may serve as good FSDH models. In fact, scientists have already started using pigs as models for Duchenne muscular dystrophy, therefore, these methods of genetically engineering pigs could easily be adapted to establish models of FSHD as well. The Tapscott Lab has demonstrated how these DUXC containing mammals will likely make great FSHD models, but do these types of animals even get muscular dystrophy? Currently, it has not been found that dogs or pigs, for example, get FSHD muscular dystrophy, but this unknown might reflect inadequacies of testing them for FSHD rather than the likelihood of them developing the disease. Dr. Tapscott explains that one of the most common early tests used to diagnose FSHD in humans is asking a person to whistle, since the facial muscles are often some of the first to experience impairment. He then adds, but “dogs don’t whistle.”

This work was funded by the National Institutes of Health, the National Institute of Arthritis and Musculoskeletal and Skin Diseases, Friends of FSH Research and the Chris Carrino Foundation.

UW/Fred Hutch Cancer Consortium member Stephen Tapscott contributed to this work.

Wong CJ, Whiddon JL, Langford AT, Belleville AE, Tapscott SJ. Canine DUXC: Implications for DUX4 retrotransposition and preclinical models of FSHD. Hum Mol Genet. 2021 Dec 9:ddab352. doi: 10.1093/hmg/ddab352. Epub ahead of print. PMID: 34888646.