Genetic diseases in the omics era

Main text

Accurate quantification of gene expression at RNA and protein levels has provided invaluable insight into the mechanisms of biological processes in health and disease. The so-called “-omics” technologies (transcriptomics, proteomics, but also epigenomics, etc.) enabled researchers to detect expression changes specifically associated with given biological states and, more recently, to establish their dynamics at a single-cell resolution in stem cell-based models. In this issue of Molecular Therapy, Djeddi et al.¹ report a meta-analysis of multi-omic data to define a common molecular signature between three forms of centronuclear myopathy (CNM), a subfamily of inherited muscle disease. They highlighted patterns of gene dysregulation associated with each mutation in well-defined mouse models, extracted the lists of common genes and ontologies, and used this information to discover new biomarkers that respond to disease correction. The study demonstrates how the power of -omics technologies can shed light on complex processes of disease initiation, progression, and correction at the molecular level.

CNMs are a group of inherited muscle disorders sharing similar clinical features and histological hallmarks. Patients suffer from muscle weakness and hypotonia, which evolves more or less severely depending on the mutation.² Muscle fibers typically show a hypotrophic phenotype with nuclei abnormally located in the center of the cell instead of at the periphery. Several genes have been associated with CNMs and are linked with specific modes of inheritance, including MTM1 (X-linked myotubular myopathy, XLMTM), DNM2 (autosomal dominant), and BIN1 (autosomal recessive). These genes establish a complex interplay in skeletal muscle cells, as MTM1 deficiency can be compensated either by DNM2 downregulation or BIN1 overexpression, suggesting that all three proteins are involved in common biological processes. The myotubularin 1 (MTM1), dynamin 2 (DNM2), and amphiphysin 2 (BIN1) proteins are known to participate in complex membrane remodeling processes between the sarcolemma, transverse tubules, and the endomembrane system. However, the pathological mechanisms of CNMs remain unclear at cellular and molecular levels.

Several preclinical therapeutic strategies have been validated in small animal models of CNMs, including the mice employed in the Djeddi et al.¹ study and in a larger Labrador model of XLMTM.^3,4 The latter was recently employed in a dose-finding study for the clinical translation of a recombinant adeno-associated virus (rAAV) expressing therapeutic MTM1 in XLMTM patients.⁵ In parallel, we published a detailed description of the muscle transcriptome in affected dogs and how it was corrected after gene transfer by inventing specific RNA-based metrics and complementary visualization tools.⁶ We showed that rAAV8-MTM1 gene transfer led to molecular rescue of the transcriptome in two muscle types and that it has the potential to predict treatment efficiency and identify potential biomarkers.

Djeddi et al.¹ built on and extended this methodology to compare different CNMs in different species so as to extract the “common molecular signature” (i.e., specific lists of differentially expressed genes [DEGs] detected in all conditions) (Figure 1A). For instance, they compared XLMTM manifestations on the transcriptome in Mtm1^−/y mice with those previously found in dogs⁶ and patients.⁷ This resulted in five common DEGs, among which the acetylcholine receptor subunits CHRNA1, CHRND, and the myogenic regulatory factor MYOG, are particularly relevant to the pathology. Importantly, the identification of these “core XLMTM genes” in an unbiased manner gives further credit to the data and the new biology discovered by the authors. Next, they conducted a rigorous comparison of the muscle transcriptome in their mouse models of CNM carrying Mtm1, Bin1, or Dnm2 mutations and highlighted 155 genes that were dysregulated in all cohorts. Interestingly, the gene ontology analysis performed on this list suggested that muscles from CNM mouse models suffer from inflammation and macrophage infiltration, which was further confirmed by immunofluorescence experiments. This unequivocally shows the interest of transcriptomics to move beyond our current understanding of pathophysiological mechanisms.

Figure 1.

Overlap of transcriptomic data helps define common molecular signatures of disease development and rescue

(A) Theoretical Venn diagram obtained after comparison of the genes dysregulated in three different centronuclear myopathies (CNMs) that highlights the “core disease signature.” (B) Venn diagram obtained after comparison of the genes rescued by three disease correction strategies and the resulting “core mechanisms of disease rescue.”

Multi-omics can also bring precious knowledge on how diseases are rescued at the molecular level. We previously introduced a battery of new metrics based on RNA-sequencing data to define genes with a “rescued” expression profile from those “resistant” to the treatment, “worsened,” or “partially rescued.”⁶ In the present study, Djeddi et al.¹ extended this concept to define the specific CNM rescue profiles in mice. Different modes of disease correction were compared, from genetic crosses to a pharmacological treatment with tamoxifen or antisense oligonucleotides. Here, it can be argued that the constitutive expression of a transgene resulting from a genetic cross does not represent a viable therapeutic strategy and remains hard to compare with postnatal interventions such as tamoxifen treatment or MTM1 gene therapy (not included in this study). However, it might also help unveil core mechanisms of disease rescue independent from the correction strategy (Figure 1B).

The authors were able to rank therapeutic strategies based on the degree of transcriptome rescue and identified Bin1 transgenic overexpression as the most efficient. How much of this is recapitulated by Mtm1 gene therapy was not investigated but could be an interesting follow-up study. The authors detected 42 genes commonly rescued by all correction strategies that were subsequently tested for their potential to serve as new therapeutic targets and circulating biomarkers. For the latter, proteomic data obtained by mass spectrometry on murine serum samples, but also from public databases, were integrated into the analysis, and the most promising candidates were further validated by conventional molecular biology techniques, including ELISA and western blot. Overall, this helped identify annexin A2 (ANXA2) as a circulating biomarker common to several CNM forms that responded to disease correction. This gene was not identified as a biomarker candidate in our previous XLMTM dog study, which did not include proteomics and complementary ELISA experiments. This suggests that inter-species differences can be a source of variability, and comparison of biomarker lists between different models might be a powerful way to isolate the most promising genes to be used to monitor disease progression and rescue in patients.

The study by Djeddi et al.¹ offers global gene expression profiling for CNMs by combining multi-omic data analysis with database mining and downstream validation. This is one convincing example of how -omics technologies can be used to improve our understanding of genetic diseases, both their progression and their correction, and how they are perceived in the community. Mutations in specific loci have broad consequences on a cell’s state and isolated phenotypes must be considered in this larger picture. Further emphasis should be placed on data from diversified sources (e.g., the epigenome, the interactome, or the metabolome) and different species. Future efforts should aim to better integrate this data using, for example, machine learning, to fully appreciate their complexity. In addition, the development of more versatile computer analysis pipelines and even better visualization tools should be encouraged. Altogether, this will help researchers and clinicians build more faithful disease models and accelerate clinical translation of the most promising treatments.