UC Davis

Morphology of the skeletal muscle is highly indicative of its function and health. In diseases like Duchenne muscular dystrophy (DMD), fibrosis develops within the muscles and the morphology of the muscle changes considerably. Histology is the gold standard for assessing the morphology of the muscles in both research and clinical settings. Whole muscles can be composed of thousands of fibers in parallel making quantification of parameters like number and size of muscle fibers difficult by manual methods. We present SMASH 2.0 as a tool for automated analysis of histological muscle images that makes use of machine learning algorithms. The classifier trained to distinguish muscle fibers in the images was able to identify muscle fibers with 99.5% sensitivity and 98.4% specificity. The data output by SMASH was not found to be significantly different from manual analysis.Skeletal muscle is also a heterogeneous tissue containing many types of cells in addition to muscle cells. Among these are satellite cells, a population of stem cells that reside at the periphery of muscle fibers and are the primary mediators of the muscle’s regenerative capabilities. Satellite cells that migrate through fibrotic tissue have been shown to have DNA damage and impaired differentiation. Methods of imaging satellite cell migration through constrictions are limited. We present two microfluidic devices, one with pores of a defined size and one that accommodates tunable collagen matrices, which are optimized for real time imaging of myoblast migration through small constrictions. These tools will help elucidate the link between constricted migration and impaired regeneration seen in DMD and can provide a method for assessing the efficacy of therapeutic treatments for muscle fibrosis.