OBJECTIVE: Alzheimer's disease (AD) is a progressive neurodegenerative disease of the central nervous system (CNS). Recently, an increased interest in the role diet plays in the pathology of AD has resulted in a focus on the detrimental effects of diets high in cholesterol and fat and the beneficial effects of caloric restriction. The current study examines how dietary composition modulates cerebral amyloidosis and neuronal integrity in the TgCRND8 mouse model of AD. METHODS: From 4 wks until 18 wks of age, male and female TgCRND8 mice were maintained on one of four diets: (1) reference (regular) commercial chow; (2) high fat/low carbohydrate custom chow (60 kcal% fat/30 kcal% protein/10 kcal% carbohydrate); (3) high protein/low carbohydrate custom chow (60 kcal% protein/30 kcal% fat/10 kcal% carbohydrate); or (4) high carbohydrate/low fat custom chow (60 kcal% carbohydrate/30 kcal% protein/10 kcal% fat). At age 18 wks, mice were sacrificed, and brains studied for (a) wet weight; (b) solubilizable Abeta content by ELISA; (c) amyloid plaque burden; (d) stereologic analysis of selected hippocampal subregions. RESULTS: Animals receiving a high fat diet showed increased brain levels of solubilizable Abeta, although we detected no effect on plaque burden. Unexpectedly, brains of mice fed a high protein/low carbohydrate diet were 5% lower in weight than brains from all other mice. In an effort to identify regions that might link loss of brain mass to cognitive function, we studied neuronal density and volume in hippocampal subregions. Neuronal density and volume in the hippocampal CA3 region of TgCRND8 mice tended to be lower in TgCRND8 mice receiving the high protein/low carbohydrate diet than in those receiving the regular chow. Neuronal density and volume were preserved in CA1 and in the dentate gyrus. INTERPRETATION: Dissociation of Abeta changes from brain mass changes raises the possibility that diet plays a role not only in modulating amyloidosis but also in modulating neuronal vulnerability. However, in the absence of a study of the effects of a high protein/low carbohydrate diet on nontransgenic mice, one cannot be certain how much, if any, of the loss of brain mass exhibited by high protein/low carbohydrate diet-fed TgCRND8 mice was due to an interaction between cerebral amyloidosis and diet. Given the recent evidence that certain factors favor the maintenance of cognitive function in the face of substantial structural neuropathology, we propose that there might also exist factors that sensitize brain neurons to some forms of neurotoxicity, including, perhaps, amyloid neurotoxicity. Identification of these factors could help reconcile the poor clinicopathological correlation between cognitive status and structural neuropathology, including amyloid pathology.