Amyloid beta (Abeta) immunoreactivity in neurons was examined in brains of 32 control subjects, 31 people with Down syndrome, and 36 patients with sporadic Alzheimer's disease to determine if intraneuronal Abeta immunoreactivity is an early manifestation of Alzheimer-type pathology leading to fibrillar plaque formation and/or neurofibrillary degeneration. The appearance of Abeta immunoreactivity in neurons in infants and stable neuron-type specific Abeta immunoreactivity in a majority of brain structures during late childhood, adulthood, and normal aging does not support this hypothesis. The absence or detection of only traces of reaction with antibodies against 4-13 aa and 8-17 aa of Abeta in neurons indicated that intraneuronal Abeta was mainly a product of alpha- and gamma-secretases (Abeta(17-40/42)). The presence of N-terminally truncated Abeta(17-40) and Abeta(17-42) in the control brains was confirmed by Western blotting and the identity of Abeta(17-40) was confirmed by mass spectrometry. The prevalence of products of alpha- and gamma -secretases in neurons and beta- and gamma-secretases in plaques argues against major contribution of Abeta-immunopositive material detected in neuronal soma to amyloid deposit in plaques. The strongest intraneuronal Abeta(17-42) immunoreactivity was observed in structures with low susceptibility to fibrillar Abeta deposition, neurofibrillary degeneration, and neuronal loss compared to areas more vulnerable to Alzheimer-type pathology. These observations indicate that the intraneuronal Abeta immunoreactivity detected in this study is not a predictor of brain amyloidosis or neurofibrillary degeneration. The constant level of Abeta immunoreactivity in structures free from neuronal pathology during essentially the entire life span suggests that intraneuronal amino-terminally truncated Abeta represents a product of normal neuronal metabolism.