Transcription factors regulate the expression of gene networks by binding to specific DNA sequences to control chromatin and transcription. These “master regulators” of the genome control cell type differentiation, define patterns during development, and modulate specific signaling pathways. Predictably, mutations in transcription factors or their binding sites underlie many human diseases. Expression of the transcription factor DUX4 in skeletal muscle –where is normally silenced– causes facioscapulohumeral muscular dystrophy (FSHD), a disease generally characterized by muscle weakness and atrophy that starts in the face (facio-) and shoulders (scapula-), progresses to the upper arms (humeral), and eventually reaches the muscles of the trunk and lower extremities. Drugs that inhibit DUX4 expression have completed advanced clinical trials or have been approved for therapeutic use in FSHD patients. However, one of the limitations of clinical trial design is the inability to locate muscles with active disease or DUX4 expression because disease severity varies within an individual’s muscles and individuals can even remain asymptomatic. Therefore, the development of methods that assess treatment efficacy at the gene-expression-level are critical to determine the beneficial results of clinical trials.
To address this need, the Tapscott lab (Human Biology) conducted a study aimed to characterize the longitudinal gene expression signature of FSHD-affected muscles in a cohort of FSHD patients. This study is a one-year follow-up of a previous report that describes the effective use of MRI to identify muscles with active disease in FSHD patients; by correlating the MRI disease score and DUX4-target gene expression levels, the researchers found that high MRI disease scores positively correlate with higher levels of DUX4-target gene expression compared to controls. Importantly, the study also identified an additional set of genes associated with extracellular matrix inflammation and complement activation as potential markers of FSHD muscle biopsies. To determine if these markers could be used as a biomarkers of disease in clinical trials, the researchers repeated muscles biopsies and performed RNA-sequencing to identify genes that separate FSHD samples from control samples. The group recently published their findings in Human Molecular Genetics.
To discriminate FSHD samples from the control samples, the researchers performed principal component analysis of the 500 most variable genes from the RNA-sequencing data. This analysis showed that FSHD samples separated from the controls and that the variance observed was not only driven by DUX4-target gene expression but also by genes associated with immune/inflammatory response, immunoglobulins, and extracellular matrix. This is an important finding because utilizing genes that are DUX4-independent as biomarkers could be useful to monitor efficacy of drugs that do not target DUX4 expression. Dr. Chao-Jen Wong, bioinformatics analyst and first author in the study, highlighted the significance of the study: “The significant contribution of this work is the identification of biomarkers of disease activity that, if validated, will provide strong basis for clinical trial design and important measures of post-therapeutic response.”
The authors then combined the RNA-sequencing expression data from the first and second visit muscle samples to (1) better identify differences that might separate FSHD samples from controls and (2) identify differences between FSHD samples from the initial evaluation and the one-year follow-up. Principal component analysis applied to k-mean clustering showed five distinct groupings that correlated with candidate biomarker expression and disease severity from MRI pathology scores. This grouping resulted in five distinct classes of gene expression patterns with variable disease severity from “mild” to “high”. Interestingly, the authors found that the mild class of FSHD samples cluster with controls. To improve the distinction capacity between mild samples and controls, the group used DESeq2 analysis to identify enriched genes in the high class that were also differentially expressed in the mild class. They identified 164 genes as potential candidates for early muscle changes in FHD that separate mild cases from controls, including DUX4 target genes and genes associated with extracellular matrix organization, cell cycle, complement activation and subsets of immune/inflammatory response genes.
The authors conclude that the detection of gene signatures associated with non-muscle cells suggests cell infiltration in FSHD-affected muscle. Dr. Wong explains the implications of this finding: “The question of cells of origin arises as the presence of cell cycle inhibitor, T and B cells suggests cell infiltration in FSHD affected muscle. Further exploration on the origin of these transcripts using single-cell RNA sequencing is crucial to establish a chronological immune-cell phenotyping.” Moving forward, the group aims to individually validate promising candidates to utilize them as reliable markers for disease progression and treatment efficacy.
This work was supported by the National Institutes of Health and the Friends of FSH Research.
Fred Hutch/UW Cancer Consortium member Dr. Tapscott contributed to this research.
Wong CJ, Wang LH, Friedman SD, et al. Longitudinal measures of RNA expression and disease activity in FSHD muscle biopsies [published online ahead of print, 2020 Feb 21]. Hum Mol Genet. 2020;ddaa031. doi:10.1093/hmg/ddaa031