FSHD myoblasts are powerless when faced with stress

From the Bradley lab, Public Health Science Division, and the Tapscott lab, Human Biology Division

Facioscapulohumeral muscular dystrophy (FSHD) is a rare muscular disorder that results from a genetic mutation that leads to a shorter chromosome 4. This is a degenerative disease in which muscles of the face, the shoulder blades and the upper arms progressively weaken until atrophy. The first symptoms are generally detected before the age of 20. Although patients suffering from FSHD have a normal lifespan, the disease is profoundly disabling, and no treatment is currently available. A decade ago, researchers have found that DUX4, a transcription activator that is normally activated during embryonic development, is aberrantly expressed in the pathologic muscle of adults, hence an important focus on the genes regulated by DUX4 in the studies of the subsequent years. However, the results published so far have mostly concentrated on RNA levels without assessing the corresponding protein levels, which can differ due to post transcriptional modifications such as translational activity or protein degradation. Researchers from Bradley and Tapscott laboratories (Public Health Science Division and Human Biology Division) tackled this unresolved issue in a recent publication in the journal Elife.

Quantitative proteomics is a powerful tool that Dr. Jagannathan and colleagues used to correlate relative amounts of protein to RNAseq data they previously published. They conducted Stable Isotope Labeling with Amino acids in Cell culture (SILAC)-based mass spectrometry on myoblast cell lines overexpressing DUX4, a model they previously characterized as faithfully mimicking the transcriptional program of FSHD cells. To do so, they cultured myoblast cells with isotope-light or isotope-heavy amino-acids and allowed the cells to integrate these labeled amino acids into their proteins. While isotope-light cells were transduced with a control vector, isotype-heavy cells were transduced with a vector enabling the inducible expression of DUX4 upon doxycycline treatment. After doxycycline treatment, both isotope-light and isotope-heavy cells were collected, pooled and processed for mass spectrometry analysis. Using this method, identical proteins can be identified while still discriminating between isotope-light versus isotope-heavy loaded proteins. The ratio of the intensity between the two isotope specific signals represents the relative quantity of each protein in the DUX4 expressing cells compared to control cells. After validating the quantitative method with known DUX4 target genes, the researchers quantified 4005 proteins from the myoblasts, for which 3961 had a corresponding RNAseq quantification. By comparing RNAseq profiles and SILAC-based mass spectrometry data, and performing a gene ontology analysis for each dataset, they discovered an immense gap between the two datasets: the RNA and proteomic profiles were drastically different. Whereas the RNA profile identified mRNA processing and splicing as the main pathways altered in DUX4 expressing myoblasts, the proteomic analysis revealed that humoral immune response, proteolysis and exocytosis were the most important processes altered in the disease. Exocytosis was of particular interest since it had never been mentioned in any previous report. Curious to validate this observation, the authors stained DUX4 transduced cells for Golgi markers and demonstrated that the Golgi apparatus was dramatically fragmented compared to control cells. This not only supported the strong relevance of their proteomic analysis, but also provided a new insight into FSHD biology, since the cellular secretory pathway is severely altered.

In continuity with this analysis, the authors paid particular attention to genes whose transcripts were induced in DUX4 expressing cells whereas protein expression was not increased. This would correspond to a post-transcriptional buffering of the DUX4-dependent reprogramming. Interestingly, Dr. Jagannathan and colleagues found that after DUX4 expression, unfolded protein response and double strand RNA stress pathways were upregulated at a transcript level without a corresponding upregulation of protein amount. This was due to the simultaneous phosphorylation of eIF2a that leads to translation inhibition. Thus, the post-transcriptional buffering may explain the absence of an efficient stress response following DUX4 overexpression.

Conversely, genes that are not transcriptionally altered by DUX4 but are downregulated at the protein level may reveal selective protein degradation or translation downregulation processes. Indeed, the researchers identified the nonsense-mediated RNA decay (NMD) pathway as inhibited at a protein level without transcript modulation, including UPF1. This pathway allows for the elimination of nonsense RNA transcripts whose translation would otherwise compromise normal cell function. The downregulation of this pathway may partially explain the degenerative aspect of the disease. This was correlated with a change in the expression of ubiquitin proteasome regulators, suggesting that the observed post-transcriptional regulation of the NMD pathway is likely due to dysregulated proteasome-mediated degradation.

Illustration showing how this study bridges a gap of knowledge between transcriptome analysis and pathology by providing quantitative proteomic data. UPF1 increased protein degradation is one of the most important findings, resulting in an inhibition of the non-sense mediated RNA decay pathway.
This study bridges a gap of knowledge between transcriptome analysis and pathology by providing quantitative proteomic data. UPF1 increased protein degradation is one of the most important findings, resulting in an inhibition of the non-sense mediated RNA decay pathway. Illustration provided by Dr. Jagannathan.

Thus, DUX4 overexpression in myoblasts not only shuts down the translation of stress response genes, but also increases the degradation of proteins involved in nonsense RNA degradation, leading to highly dysfunctional cells. As Dr. Jagannathan comments: “Our study has provided the first quantitative assessment of the DUX4-induced proteome, paving way for future mechanistic studies on how DUX4’s transcriptional targets induce the FSHD pathology. We think of this as a missing piece of an important puzzle that will allow us to finally piece together the mechanism of FSHD pathogenesis”.

This story is only the start of long process in the fight against FSHD for the Fred Hutch researchers. The next objectives are multiple, Dr. Jagannathan explains: “Our findings raise several questions: How does DUX4 induce proteasomal degradation of RNA quality control factors to cause shut down NMD? Is a similar regulation also found in the early embryo where DUX4 was recently shown to be expressed?  These are exciting directions to pursue that will lead to a better understanding of the physiological role of DUX4/NMD regulatory feedback loop and how this can be disrupted in order to develop novel therapeutics for FSHD”.


This work was supported by National Institutes of Health and the FSH society.

Cancer Consortium faculty members Drs. Robert Bradley and Stephen Tapscott contributed to this research.

Jagannathan S, Ogata Y, Gafken PR, Tapscott SJ, Bradley RK. 2019. Quantitative proteomics reveals key roles for post-transcriptional gene regulation in the molecular pathology of facioscapulohumeral muscular dystrophy. eLIfe. 8:e41740. DOI: https://doi.org/10.7554/eLife.41740