Single APP locus knock-in mouse models of Alzheimer's disease: Distinguishing facts from artifacts

Abstract

Transgenic (Tg) mice overexpressing amyloid precursor protein (APP) have made tremendous contributions to the experimental understanding of Alzheimer's disease (AD) pathophysiology. These 1st generation models however suffer from the following intrinsic problems. [1] APP is expressed to an unphysiologically high extent. Because APP interacts (indirectly) with kinesin, the axonal transport is disturbed, possibly generating artificial phenotype(s). [2] APP fragments other than Abeta, i.e. APPs, AICD and CTF-beta, are also overproduced and could also generate artifact(s). No perfect negative controls exist. [3] The transgene is inserted into a gene locus of the host animal and affect endogenous gene expression. [4] Expression of the transgene varies from line to line and from time to time. [5] APP-Tg mice often die of unknown cause(s). [6] Cross-breeding APP-Tg mice with other mutant mice is likely to generate even more complicated artifacts. Indeed, calpastatin deficiency caused overt early lethality in APP-transgenic mice. In order to overcome these drawbacks, we generated genomic knock-in (KI) mice which harbor the Swedish and Beyreuther/Iberian mutations with the Abeta sequence humanized by homologous recombination (NL-F mice). As a negative control, KI mice without the Beyreuther/Iberian mutation (NL mice) were produced. We also generated KI mice that carry the Arctic mutation in addition to the Beyreuther/Iberian mutation (NL-G-F mice). These 2 nd generation (NL-F mice) mice start accumulating Abeta at about 6 months and showed exponential deposition up to 24 months. Calpastatin deficiency accelerated Abeta pathology and cognitive dysfunction without causing early lethality. Our results indicate that the early lethality observed in the calpastatin-deficient APP-Tg mice is likely an experimental artifact and that the calpain-calpastatin system plays a mechanistic role in the pathophysiology of AD. Another line of knockin mice carrying the Arctic mutation (NL-G-F mice) displayed more aggressive Abeta deposition not only in cortex but also in subcortical regions, which may represent characteristic pathology of this particular form of familial AD. Our KI mice will be a valuable rule for identifying the mechanisms that lead to tauopathy and neurodegeneration with minimum artifact. We propose that the knockin mice be used for global collaboration for identification of therapeutic targets.