A Convolutional Neural Network for the automatic diagnosis of collagen VI-related muscular dystrophies

Personalized in vitro Extracellular Matrix Models of Collagen VI-Related Muscular Dystrophies

Collagen VI-related dystrophies (COL6-RDs) are a group of rare congenital neuromuscular dystrophies that represent a continuum of overlapping clinical phenotypes that go from the milder Bethlem myopathy (BM) to the severe Ullrich congenital muscular dystrophy, for which there is no effective treatment. Mutations in one of the three Collagen VI genes alter the incorporation of this protein into the extracellular matrix (ECM), affecting the assembly and the structural integrity of the whole fibrillar network. Clinical hallmarks of COL6-RDs are secondary to the ECM disruption and include muscle weakness, proximal joint contractures, and distal hyperlaxity. Although some traits have been identified in patients’ ECMs, a correlation between the ECM features and the clinical phenotype has not been established, mainly due to the lack of predictive and reliable models of the pathology. Herein, we engineered a new personalized pre-clinical model of COL6-RDs using cell-derived matrices (CDMs) technology to better recapitulate the complexity of the native scenario. We found that CDMs from COL6-RD patients presented alterations in ECM structure and composition, showing a significantly decreased Collagen VI secretion, especially in the more severe phenotypes, and a decrease in Fibrillin-1 inclusion. Next, we examined the Collagen VI-mediated deposition of Fibronectin in the ECM, finding a higher alignment, length, width, and straightness than in patients with COL6-RDs. Overall, these results indicate that CDMs models are promising tools to explore the alterations that arise in the composition and fibrillar architecture due to mutations in Collagen VI genes, especially in early stages of matrix organization. Ultimately, CDMs derived from COL6-RD patients may become relevant pre-clinical models, which may help identifying novel biomarkers to be employed in the clinics and to investigate novel therapeutic targets and treatments.

CRISPR/Cas9-Mediated Allele-Specific Disruption of a Dominant COL6A1 Pathogenic Variant Improves Collagen VI Network in Patient Fibroblasts

Collagen VI-related disorders are the second most common congenital muscular dystrophies for which no treatments are presently available. They are mostly caused by dominant-negative pathogenic variants in the genes encoding α chains of collagen VI, a heteromeric network forming collagen; for example, the c.877G>A; p.Gly293Arg COL6A1 variant, which alters the proper association of the tetramers to form microfibrils. We tested the potential of CRISPR/Cas9-based genome editing to silence or correct (using a donor template) a mutant allele in the dermal fibroblasts of four individuals bearing the c.877G>A pathogenic variant. Evaluation of gene-edited cells by next-generation sequencing revealed that correction of the mutant allele by homologous-directed repair occurred at a frequency lower than 1%. However, the presence of frameshift variants and others that provoked the silencing of the mutant allele were found in >40% of reads, with no effects on the wild-type allele. This was confirmed by droplet digital PCR with allele-specific probes, which revealed a reduction in the expression of the mutant allele. Finally, immunofluorescence analyses revealed a recovery in the collagen VI extracellular matrix. In summary, we demonstrate that CRISPR/Cas9 gene-edition can specifically reverse the pathogenic effects of a dominant negative variant in COL6A1.