The present study examines age-related changes in the subcellular localization of cathepsin B (cath B) and cathepsin D (cath D), as well as morphological features of the cathepsin-immunoreactive (ir) neurons in rat cerebral cortex. Sprague-Dawley rats were studied at 3 and 26 months. By immunoelectron microscopy cath B- or cath D-immunoreactivities were found in many, but not all, pyramidal neurons. In young rat cerebral cortical neurons, cath B was observed not only in lysosomal systems such as multivesicular bodies, dense bodies, and lipofuscin granules, but also in extralysosomal sites. By contrast, cath D was confined mainly to lysosomal systems in young rats. In aged rats, cath B showed a similar pattern in its subcellular localization compared to the young control, but some of the dense bodies containing cath B was closely apposed to the outer nuclear envelope. These cells exhibited a relatively normal appearance. Regardless of subcellular localization, approximately 10% of cath B-ir neurons displayed ultrastructural disturbances presumed to indicate an early stage of degeneration. The nucleus was indented, nuclear boundary was indistinct, nuclear pore structures appeared separately with high frequency, and the endoplasmic reticulum appeared to be affected. In addition to its presence in lysosomal structures, cath D-immunoreactivity in aged cerebral cortex was noted prominently in the cytosol as diffuse granules. About 37% of cath D-ir cells showed this age-related change. Among the neurons with the diffusely scattered form of cath D, approximately 70% of cells exhibited the degenerating features. These cells were characterized by large amounts of diffuse cath D, reduced cellular size, loss of the nuclear boundary, scattered nuclear pore structures, an often fragmentation of the nucleus, disturbances of endoplasmic reticular system, and in advanced stages, condensed nucleus and poor preservation of almost cytoplasmic organelles. Though some of these features were also found in cath B-ir neurons, findings of overt degeneration, such as fragmented and condensed nuclei and impaired almost cytoplasmic organelles, were generally not observed in cath B-ir neurons. In addition, lipofuscin aggregates containing cath D were observed frequently in the extracellular space close to sites of ruptured plasma membrane, whereas in the sections stained with anti-cath B antibodies, large-sized lipofuscin aggregates showed only very weak or no cath B-immunoreactivity at all. Taken together, the present results suggest that cath D and cath B may be regulated differently and play their specific roles in the aging of the brain, especially, the change in location of cath D from the lysosomal system to the cytosol in the aged brain may play an important role in age-related cell death.