Spinocerebellar ataxia type 2 (SCA2) is an autosomal dominantly inherited, neurodegenerative disease. It can manifest either with a cerebellar syndrome or as Parkinson's syndrome, while later stages involve mainly brainstem, spinal cord and thalamus. This particular atrophy pattern resembles sporadic multi-system-atrophy (MSA) and results in some clinical features indicative of SCA2, such as early saccade slowing, early hyporeflexia, severe tremor of postural or action type, and early myoclonus. For treatment, levodopa is temporarily useful for rigidity/bradykinesia and for tremor, magnesium for muscle cramps, but neuroprotective therapy will depend on the elucidation of pathogenesis. The disease cause lies in the polyglutamine domain of the protein ataxin-2, which can expand in families over successive generations resulting in earlier onset age and faster progression. Genetic testing in SCA2 and other polyglutamine disorders like the well-studied Huntington's disease is now readily available for family planning. Although these disorders differ clinically and in the affected neuron populations, it is not understood how the different polyglutamine proteins mediate such tissue specificity. The neuronal intranuclear inclusion bodies described in other polyglutamine disorders are not frequent in SCA2. For the quite ubiquitously expressed ataxin-2, a subcellular localization at the Golgi, the endoplasmic reticulum and the plasma membrane, in interaction with proteins of mRNA translation and of endocytosis have been observed. As a first victim of SCA2 degeneration, cerebellar Purkinje neurons may be preferentially susceptible to alterations of these subcellular pathways, and therefore our review aims to portray the particular profile of the SCA2 disease process and correlate it to the specific features of ataxin-2.