Unstable expansions in the Q22-polyglutamine domain of human ATXN2 mediate risks for motor neuron diseases such as ALS/FTLD or cause the autosomal dominant Spinocerebellar Ataxia type 2 (SCA2), but the pathogenesis is not understood and models are unavailable. We generated a novel knock-in mouse line with CAG100 expansion in Atxn2, transmitted unstably. The mutant protein accumulated in neuronal cytosolic aggregates, with a characteristic pattern of multi-system-atrophy. Loss-of-function phenotypes included less mutant offspring, initial weight gain and motor hyperactivity. Progressive toxic aggregation effects started around 20 weeks in homozygous animals showing weight loss, reduced muscle strength and gait ataxia. Lifespan was decreased. In the cerebellum, neuronal soma and dendrites were remarkably spared. However, myelin proteins MBP, CNP, PLP1 and transcripts Mal, Mobp, Rtn4 decreased markedly, especially adhesion factors MAG and MOG. In neurons, strong reductions were found for mRNAs of perineuronal elements Hapln1, Hapln2, Hapln4, of axonal myelin interactors Prnp and Klk6. At protein level, the adhesion factor neuroplastin and neurofilaments were strongly reduced, while presynaptic alpha-synuclein increased two-fold. Overall, this authentic SCA2 mouse model elucidates how altered function and aggregation toxicity of ATXN2 conspire to trigger axon-myelin disconnection. This model will promote the development of neuroprotective therapies and disease biomarkers.

Atxn2-CAG100-knock-in affects mouse lifespan and vestibulo-cerebellar function via neural disconnection