Neurodegeneration in Alzheimer’s disease (AD) is extensively studied, and the involvement of astrocytes and other cell types in this process has been described. However, the responses of astrocytes themselves to amyloid β peptides ((Aβ; the widely accepted major toxic factor in AD) is less well understood. Here, we show that Aβ(1-42) is toxic to primary cultures of astrocytes. Toxicity does not involve disruption of astrocyte Ca2+ homeostasis, but instead occurs via formation of the toxic reactive species, peroxynitrite. Thus, Aβ(1-42) raises peroxynitrite levels in astrocytes, and Aβ(1-42) toxicity can be inhibited by antioxidants, or by inhibition of nitric oxide (NO) formation (reactive oxygen species (ROS) and NO combine to form peroxynitrite), or by a scavenger of peroxynitrite. Increased ROS levels observed following Aβ(1-42) application were derived from NADPH oxidase. Induction of haem oxygenase-1 (HO-1) protected astrocytes from Aβ(1-42) toxicity, and this protective effect was mimicked by application of the carbon monoxide (CO) releasing molecule CORM-2, suggesting HO-1 protection was attributable to its formation of CO. CO suppressed the rise of NADPH oxidase-derived ROS caused by Aβ(1-42). Under hypoxic conditions (0.5% O2, 48 h) HO-1 was induced in astrocytes and Aβ(1-42) toxicity was significantly reduced, an effect which was reversed by the specific HO-1 inhibitor, QC-15. Our data suggest that Aβ(1-42) is toxic to astrocytes, but that induction of HO-1 affords protection against this toxicity due to formation of CO. HO-1 induction, or CO donors, would appear to present attractive possible approaches to provide protection of both neuronal and non-neuronal cell types from the degenerative effects of AD in the central nervous system.