Glucocorticoid modulation of androgen receptor nuclear aggregation and cellular toxicity is associated with distinct forms of soluble expanded polyglutamine protein.

Spinobulbar muscular atrophy is a progressive motor neuron disease caused by abnormal polyglutamine tract expansion in the androgen receptor (AR) gene, and is part of a family of central nervous system (CNS) neurodegenerative diseases, including Huntington's disease (HD). Each pathologic protein is widely expressed, but the cause of neuronal degeneration within the CNS remains unknown. Many reports now link abnormal polyglutamine protein aggregation to pathogenesis. A previous study reported that activation of the wild-type glucocorticoid receptor (wtGR) suppressed the aggregation of expanded polyglutamine proteins derived from AR and huntingtin, whereas a mutant receptor containing an internal deletion, GRDelta108-317, increased polyglutamine protein aggregation, in this case primarily within the nucleus. In this study, we use these two forms of GR to study expanded polyglutamine AR protein in different cell contexts. Using cell biology and biochemical approaches, we find that wtGR promotes soluble forms of the protein and prevents nuclear aggregation in NIH3T3 cells and cultured neurons. In contrast, GRDelta108-317 decreases polyglutamine protein solubility, and causes formation of nuclear aggregates in non-neuronal cells. Nuclear aggregates recruit hsp72 more rapidly than cytoplasmic aggregates, and are associated with decreased cell viability. Limited proteolysis and chemical cross-linking suggest unique soluble forms of the expanded AR protein underlie these distinct biological activities. These observations provide an experimental framework to understand why expanded polyglutamine proteins may be toxic only to certain populations of cells, and suggest that unique protein associations or conformations of expanded polyglutamine proteins may determine subsequent cellular effects such as nuclear localization and cellular toxicity.