The fibrillar beta-amyloid protein (A beta) plaques of Alzheimer's disease (AD) are associated with reactive astrocytes and dystrophic neurites and have been suggested to contribute to neurodegenerative events in the disease. We recently reported parallel in vitro and in situ findings, suggesting that the adoption of a reactive phenotype and the colocalization of astrocytes with plaques in AD may be mediated in large part by aggregated A beta. Thus, A beta-mediated effects on astrocytes may directly affect disease progression by modifying the degenerative plaque environment. Alternatively, plaque-associated reactive astrocytosis may primarily represent a glial response to the neural injury associated with plaques and not significantly contribute to AD pathology. To investigate the validity of these two positions, we examined the differential colocalization of reactive astrocytes and dystrophic neurites with plaques. Hippocampal sections from AD brains--ranging in neuropathology from mild to severe--were triple-labeled with antibodies recognizing A beta protein, reactive astrocytes, and dystrophic neurites. We observed not only plaques containing both or neither cell type, but also plaques containing (1) reactive astrocytes but not dystrophic neurites and (2) dystrophic neurites but not reactive astrocytes. The relative proportion of plaques colocalized with reactive astrocytes in the absence of dystrophic neurites is relatively high in mild AD but significantly decreases over the course of the disease, suggesting that plaque-associated astrocytosis may be an early and perhaps contributory event in AD pathology rather than merely a response to neuronal injury. These data underscore the potentially significant contributions of reactive astrocytosis in modifying the plaque environment in particular and disease progression in general.