Alexander M. Rossor, Emily C. Oates, Hannah K. Salter, Yang Liu, Sinead M. Murphy, Rebecca Schule, Michael A. Gonzalez, Mariacristina Scoto, Rahul Phadke, Caroline A. Sewry, Henry Houlden, Albena Jordanova, Iyailo Tournev, Teodora Chamova, Ivan Litvinenko, Stephan Zuchner, David N. Herrmann, Julian Blake, Janet E. Sowden, Gyuda Acsadi, Michael L. Rodriguez, Manoj P. Menezes, Nigel F. Clarke, Michaela Auer Grumbach, Simon L. Bullock, Francesco Muntoni, Mary M. Reilly, Kathryn N. North

Spinal muscular atrophy is a disorder of lower motor neurons, most commonly caused by recessive mutations in SMN1 on chromosome 5q. Cases without SMN1 mutations are subclassified according to phenotype. Spinal muscular atrophy, lower extremity-predominant, is characterized by lower limb muscle weakness and wasting, associated with reduced numbers of lumbar motor neurons and is caused by mutations in DYNC1H1, which encodes a microtubule motor protein in the dynein-dynactin complex and one of its cargo adaptors, BICD2. We have now identified 32 patients with BICD2 mutations from nine different families, providing detailed insights into the clinical phenotype and natural history of BICD2 disease. BICD2 spinal muscular atrophy, lower extremity predominant most commonly presents with delayed motor milestones and ankle contractures. Additional features at presentation include arthrogryposis and congenital dislocation of the hips. In all affected individuals, weakness and wasting is lower-limb predominant, and typically involves both proximal and distal muscle groups. There is no evidence of sensory nerve involvement. Upper motor neuron signs are a prominent feature in a subset of individuals, including one family with exclusively adult-onset upper motor neuron features, consistent with a diagnosis of hereditary spastic paraplegia. In all cohort members, lower motor neuron features were static or only slowly progressive, and the majority remained ambulant throughout life. Muscle MRI in six individuals showed a common pattern of muscle involvement with fat deposition in most thigh muscles, but sparing of the adductors and semitendinosus. Muscle pathology findings were highly variable and included pseudomyopathic features, neuropathic features, and minimal change. The six causative mutations, including one not previously reported, result in amino acid changes within all three coiled-coil domains of the BICD2 protein, and include a possible 'hot spot' mutation, p.Ser107Leu present in four families. We used the recently solved crystal structure of a highly conserved region of the Drosophila orthologue of BICD2 to further-explore how the p.Glu774Gly substitution inhibits the binding of BICD2 to Rab6. Overall, the features of BICD2 spinal muscular atrophy, lower extremity predominant are consistent with a pathological process that preferentially affects lumbar lower motor neurons, with or without additional upper motor neuron involvement. Defining the phenotypic features in this, the largest BICD2 disease cohort reported to date, will facilitate focused genetic testing and filtering of next generation sequencing-derived variants in cases with similar features.