The key diagnostic feature of the neuropathology in Alexander disease is the widespread deposition of Rosenthal fibers in subpial, periventricular, and white-matter astrocytes throughout the CNS. Previous transgenic results show that mice are capable of building the biochemically complex Rosenthal fibers within a short period of time in vivo. These transgenic lines also showed a significant phenotype, either early death or low body weight, indicating that they are sensitive to at least some forms of astrocyte dysfunction. The existing models are thus useful for examining several aspects of Alexander disease. However, because they do not display most of the common symptoms and pathology of the disorder, mice are now being produced with targeted point mutations in their endogenous Gfap gene. These findings can provide more complete models, leading to continued progress in unraveling the pathogenesis of Alexander disease and facilitating the development and testing of strategies for potential therapy. © 2004 Elsevier Inc. All rights reserved.