Prion protein (PrP) from the brains of animals with transmissible spongiform encephalopathies is partially protease resistant (PrP^res) compared with fully sensitive PrP (PrP^sen) from uninfected brains. In most experimental models, PrP^res is a reliable indicator of infectivity. Light microscopic studies have suggested that both PrP^sen and disease‐specific accumulations of PrP are associated with follicular dendritic cells (FDCs). Using immunogold electron microscopy, this study has demonstrated disease‐specific accumulation of PrP in the spleens of C57 BL mice, 70 days after intracerebral infection with the ME7 strain of scrapie and at the terminal stage of disease at 170 days. At both stages, tingible body macrophages contained PrP within lysosomes and PrP was also detected at the plasmalemma of FDCs. In the light zone of follicles of terminally diseased mice, all FDC dendrites were arranged in the form of highly reactive or hyperplastic labyrinthine glomerular complexes, within which PrP was consistently seen between FDC processes in association with abundant electron dense material, interpreted as antigen–antibody complexes. Within some glomeruli, fibrillar forms of PrP consistent with amyloid were seen. At 70 days after challenge, large or hyperplastic labyrinthine complexes were rare and invariably labelled for PrP. However, sparse PrP labelling was also seen on simple FDC processes at this stage. The ubiquitous accumulation of extracellular PrP in complex glomerular dendrites of FDCs in spleens from terminally affected mice, contrasted with simple FDC profiles, sparse PrP and limited electron dense deposits in all but a few FDCs of 70‐day post‐infected mice. This suggests that FDCs continually release PrP from the cell surface, where it is associated with trapped antigen–antibody complexes and dendritic extension. It is likely that tingible body macrophages acquire PrP following phagocytosis of PrP within iccosomes or from the extracellular space around FDC dendrites. These studies would not support an intracellular phase of PrP accumulation in FDCs but show that PrP is produced in excess by scrapie‐infected cells from where it is released into the extracellular space. We suggest that PrP^sen is involved in dendritic extension or in the process of antibody–antigen trapping, perhaps as part of the binding mechanism for antigen–antibody complexes. © Crown copyright 2000. Reproduced with the permission of Her Majesty's Stationery Office. Published by John Wiley & Sons, Ltd.