Centronuclear myopathies (CNM) are clinically and genetically heterogeneous disorders characterized by muscle atrophy and abnormal nuclear centralization in muscle fibers. Autosomal dominant CNM primarily results from mutations in DNM2, encoding the GTPase dynamin 2, a key regulator of membrane remodelling and endocytosis. The clinical presentation of the patients correlates with the position of the mutation, and ranges from severe neonatal hypotonia and respiratory distress to milder adult-onset forms. The most common mutation c.1393C>T (R465W) is found in 25% of all cases, and affected individuals typically show moderate generalized muscle weakness, ptosis, and ophthalmoplegia. Functional investigations suggest that the DNM2 mutations induce a gain-of-function, and the genetic or pharmacological downregulation of dynamin 2 efficiently antagonizes the disease development in the mouse model. Here we characterize a spontaneous CNM dog model, and we provide phenotypical, genetic, histological, and ultrastructural data through a longitudinal study over a period of 18 months. We identified the common DNM2 mutation R465W in a Border collie, and crossing with a healthy Beagle dam resulted in four affected offspring carrying the mutation, and a single healthy dog without the mutation. The affected dogs presented with mildly progressive muscle weakness from the age of 4 months, and MRI revealed muscle atrophy in particular of the paravertebral and masticatory muscles. Noteworthy, histological anomalies preceded the clinical signs, and the CNM-typical hallmarks myofiber atrophy, endomysial fibrosis, central nuclei, necklace fibers, and mitochondrial clustering were seen in biopsies at 2 months of age. Electron microscopy performed at 6 months of age detected mitochondrial anomalies, and functional investigations confirmed mitochondrial dysfunction. Overall, DNM2 dogs manifested a slowly progressive congenital myopathy and recapitulated the histological picture of the human disorder. The present study also provides insight into the sequence of events leading to muscle dysfunction. There is currently no treatment for DNM2-related CNM, and the canine model represents a valuable tool to validate pre-clinical treatment options on a large mammal based on proof-of-concepts recently obtained in the mouse.