Super-resolution imaging with deep learning-based segmentation for detailed characterization of mitochondrial arrangement in Pompe disease skeletal muscle

Pompe disease (glycogen storage disease type II) is an autosomal recessive lysosomal storage disorder characterized by progressive glycogen accumulation within lysosomes. It leads to their enlargement, autophagosome build-up and defective autophagic flux. Among the pathophysiological features, mitochondrial abnormalities have long been regarded as secondary consequences of lysosomal dysfunction. Typically, they have been described in electron microscopy, revealing paracrystalline inclusions, cristae lost, swollen mitochondria, and glycogen-filled structures. However, the spatial organization and interplay between mitochondria and lysosomes in skeletal muscle remain poorly understood, as does the progression of these alterations with respect to muscle metabolic profile. Here, we present a novel approach combining super-resolution imaging with a deep learning- based image analysis workflow to quantitatively assess mitochondrial and lysosomal remodeling as well as their interactions in skeletal muscle of the main murine model of the Pompe disease. Organelles were analyzed at two specific stages of the disease, according to muscle type, fiber type and subcellular location of the mitochondria. We show that the overall structure of the mitochondrial network is affected as early as the pre-symptomatic stage (1 month), while changes in mitochondrial density are more restricted at this stage and become more widespread as disease progresses (4 months). Importantly, these pathophysiological modifications are highly dependent on the muscle, fiber type and subcellular location. Alongside a rapid and widespread increase in lysosomal size, and a subsequent shift toward tighter lysosomal clustering at the later stage, we observe a progressive, region-specific increase in mitochondria-lysosome interactions that is most pronounced in the intermyofibrillar region. Our findings establish that this original imaging approach provides a relevant and powerful framework for quantitatively analyzing interactions between organelles within skeletal muscle fibers, thus offering new opportunities to explore the subcellular changes underlying disease progression. As such, it represents an interesting tool for monitoring pathophysiology and evaluating the effectiveness of therapeutic interventions.