Novel, highly translational learning assessment in a mouse model of Alzheimer's disease

Abstract

The ability to apply previously learned information to novel problems and situations has been referred to as reconfiguration of stimuli or ‘transfer learning’. Deficits in this type of learning, in humans, correlate with mild hippocampal atrophy in non-demented elderly and appear to aid in the prediction of the progression into Alzheimer's disease (AD). To date, theories and models regarding this facet of hippocampal function have not been applied to animal models of pathological aging associated with hippocampal/medial temporal lobe deficits despite the fact that such learning offers direct translational advantages compared to conventional human and rodent assessments. In the current study, a mouse task was developed to determine if “transfer learning” assessment was sensitive to pathological changes in a mouse model of AD (APPswe, PSEN1dE9)85Dbo/o; APP+PS1 mice) and if this type of learning was hippocampal dependent. In the rodent task, mice learned a series of concurrent discriminations in one dimension (pairs of odors/digging media) in the presence of a feature-irrelevant stimulus in the other dimension not predictive of reward (learning phase). This irrelevant stimulus was then changed and discrimination performance evaluated (transfer phase). Three month old APP+PS1 mice were not impaired in initial discrimination learning or on the ability to transfer this learned information to the altered context. In contrast, at 12 months of age, APP+PS1 mice learned the initial concurrent discriminations on par with NTgs, but were impaired when required to “transfer” this learning into a new configuration/context. There were no differences in spatial learning in the Morris water maze (a hippocampal dependent task) between the APP+PS1 and NTgs at 12 months of age. Furthermore, ibotenic acid was used to produce bilateral neurotoxic hippocampal lesions in a cohort of C57Bl/6J mice. Lesioned mice were not impaired in the initial discrimination learning, but were impaired when required to transfer this learned information to the new context. Also, lesioned mice were impaired in the spatial version of the Morris water maze. These data describe a novel and sensitive hippocampal-dependent paradigm (“transfer learning”) that is analogous in humans and mice, appears to be sensitive to AD-like pathology and is well-suited to within-subject experimental design.