The complement protein, C1q, has been shown to bind to fibrillar beta-amyloid, resulting in the activation of the classical complement pathway. C1q has also been found associated with most but not all amyloid deposits in brain. To determine whether C1q is exclusively associated with plaques containing the fibrillar form of beta-amyloid, normal and Alzheimer brain were immunohistochemically double labeled using thioflavine, which specifically stains beta-amyloid in a beta-sheet conformation, and an affinity- purified antibody to human C1q. C1q immunostaining was colocalized with nearly all thioflavine-positive plaques, while C1q was not detected in beta-amyloid immunopositive plaques which were thioflavine-negative. Beta-amyloid plaques in nondemented controls (which are typically thioflavine-negative) were also negative for C1q. Microglia and astrocytes of reactive morphology were also associated with C1q-positive plaques and neurons. Interestingly, many neuronal cells in the AD brain, but not microglia or astrocytes, stained prominently with anti-C1q. Neurons in control brain were not C1q positive. Our data suggest that some of these C1q-positive structures were neurofibrillary tangles immunoreactive for hyperphosphorylated tau, which may be binding extracellular C1q. However, a large number of the C1q-positive neurons had intact cell morphology; suggesting that these cells may be synthesizing this critical complement component. Since the presence of C1q suggests the activation of complement and/or the activation of proinflammatory events, and the specific class of plaques that contain C1q are the type that corresponds to observed clinical dementia, these findings further support the hypothesis that complement plays a role in the pathogenesis of AD.