The article is of interest because the J20 mice are an artificial and imperfect model for early onset family AD that combines the double swedish mutation at the BACE site with the Indiana mutation at the gamma-secrtase site. Moreover the involvement of protein tau as claimed is not verified independently

Consequently - and in general - one should be most careful to extrapolate pre-clinical data obtained in a mouse model to a clinical human setting - a point that we have advocated over and over again vàv our own mouse models for amyloid & tau pathology: "models" they are, no more - no less ...

Article Tau immunization: A cautionary tale?