Late-long-term potentiation magnitude, but not Aβ levels and amyloid pathology, is associated with behavioral performance in a rat knock-in model of Alzheimer disease

Cleavage of Amyloid precursor protein by β- and γ-secretases lead to Aβ formation. The widely accepted pathogenic model states that these mutations cause AD via an increase in Aβ formation and accumulation of Aβ in Amyloid plaques. APP mutations cause early onset familial forms of Alzheimer’s disease (FAD) in humans. We generated App−Swedish (Apps) knock−in rats, which carry a pathogenic APP mutation in the endogenous rat App gene. This mutation increases β-secretase processing of APP leading to both augmented Aβ production and facilitation of glutamate release in Apps/s rats, via a β-secretase and APP−dependent glutamate release mechanism. Here, we studied 11 to 14-month-old male and female Apps/s rats. To determine whether the Swedish App mutation leads to behavioral deficits, Apps/s knock-in rats were subjected to behavioral analysis using the IntelliCage platform, an automated behavioral testing system. This system allows behavioral assessment in socially housed animals reflecting a more natural, less stress-inducing environment and eliminates experimenter error and bias while increasing precision of measurements. Surprisingly, a spatial discrimination and flexibility task that can reveal deficits in higher order brain function showed that Apps/s females, but not Apps/s male rats, performed significantly worse than same sex controls. Moreover, female control rats performed significantly better than control and Apps/s male rats. The Swedish mutation causes a significant increase in Aβ production in 14-month-old animals of both sexes. Yet, male and female Apps/s rats showed no evidence of AD−related amyloid pathology. Finally, Apps/s rats did not show signs of significant neuroinflammation. Given that the APP Swedish mutation causes alterations in glutamate release, we analyzed Long-term potentiation (LTP), a long-lasting form of synaptic plasticity that is a cellular basis for learning and memory. Strikingly, LTP was significantly increased in Apps/s control females compared to both Apps/s sexes and control males. In conclusion, this study shows that behavioral performances are sex and App-genotype dependent. In addition, they are associated with LTP values and not Aβ or AD-related pathology. These data, and the failures of anti-Aβ therapies in humans, suggest that alternative pathways, such as those leading to LTP dysfunction, should be targeted for disease-modifying AD therapy.