Objectives: The presence of beta-amyloid (beta A) deposition, induction of reactive gliosis and dystrophic neurites, is a characteristic feature of neuritic plaques in Alzheimer's disease. In vitro, beta A-exposed astrocytes become reactive, similar to astrocytes in contact with beta A plaques in vivo. How beta A-exposed reactive astrocytes support neuron process growth, however, is not well defined. Therefore, we used neuron/astrocyte co-cultures in which astrocytes had been grown on beta A, to assess whether process growth was altered. Methods: Purified rat cortical astrocytes were plated on the beta A peptide's neurotoxic fragment (25-35), the scrambled (35-25) peptide, or poly-D-lysine alone and grown to confluency before mouse cortical neurons were seeded at low density onto the astrocyte monolayer. Cell survival was assessed using tyrpan blue, lactate dehydrogenase release and propidium iodide. Process growth was analyzed using specific antibodies against MAP2 and the 200 kDa neurofilament subunit (NF-H) to identify dendrites and axons, respectively. Results: beta A-exposed astrocytes changed dramatically from their flat polygonal shape into stellate process-bearing morphology. Viability however, was not affected. Immunocytochemical analysis of neuronal processes using anti-MAP2 and anti-NF-H, demonstrated that beta A (25-35)induced reactive astrocytes had an altered ability to support dendrite and axon growth after 3 days in vitro. Indeed, primary dendrite number and axon length were decreased by 30 and 26%, respectively, compared with control astrocytes, whereas individual primary dendrite length increased by 20%. Astrocyte support of dendritic branching, however, was not affected by beta A. Discussion: We conclude that an astrocyte reaction to beta A may contribute, in part, to neuronal dystrophy associated with beta A plaques.