Research Models

APP23 x PS1-R278I

Species: Mouse
Genes: APP, PSEN1
Mutations: PSEN1 R278I
Modification: APP: Transgenic; PSEN1: Knock-In
Disease Relevance: Alzheimer's Disease
Strain Name: B6.Cg-Tg(Thy1-APP)3Somm/J; Psen1tm1.1Tcs
Genetic Background: C57BL/6J
Availability: Available through Takaomi Saido

This mouse model is a cross between a well-characterized APP transgenic (APP23) and a PSEN1 knock-in mouse (PS1-R278I) that expresses human PSEN1 with a mutation linked to atypical AD. The R278I mutation alters γ-secretase processing of APP, leading to unusually high levels Aβ43 with correspondingly low levels of Aβ40 due to impaired trimming of the Aβ43 peptide (Nakaya et al., 2005). Consistent with these in vitro findings, PS1-R278I mice and APP23 X PS1-R278I mice exhibit high levels of Aβ43 from an early age. The presence of the PSEN1 mutation in the double mutant accelerates plaque development and memory deficits relative to APP23 mice, suggesting that the skewed ratio of Aβ peptides, and perhaps elevated Aβ43 in particular, is deleterious (Saito et al., 2011).

Heterozygous PS1-R278I mice were used for the cross because homozygous mice die in utero due to insufficient γ-secretase activity. Heterozygous animals, on the other hand, develop normally despite a 50 percent reduction in γ-secretase activity. As expected, the heterozygous mice exhibit elevated levels of Aβ43 in brain fractions, and correspondingly low Aβ40. Although the heterozygous knock-in mice appear phenotypically normal as adults, it is unknown if additional behavioral and pathological analyses would reveal subtle abnormalities.

When crossed with APP23 mice, which overexpress APP with the Swedish mutation, the resulting progeny had accelerated plaque development and memory deficits relative to APP23 littermates. The double mutant started to accumulate Aβ deposits at 6 months of age, whereas the APP23 mice in this study did not develop a commensurate plaque load until 12 months of age. The double mutant also had a higher density of thioflavin-S positive plaques, suggesting that Aβ43 may foster core formation. Around 9 months of age, large numbers of reactive astrocytes were observed around plaques and pyroglutamate Aβ (N3pE-Aβ) co-localized with plaques.

Prior to plaque formation, a short-term memory deficit was observed in APP23 X PS1–R278I mice. At 3 to 4 months of age, the double mutant performed less well than APP23 mice in the Y maze. A similar deficit was seen in the Morris water maze, but it was not statistically significant.  

Biochemical analysis showed that the brains of the double-mutant mice exhibited an early and selective increase in Aβ43 in the fraction extracted with Tris-HCl buffered saline as well as the fraction soluble in guanidine hydrochloride. Elevated Aβ43 was detectable at 3 months of age, which is prior to plaque deposition and coincident with memory impairment. Aβ40 and Aβ42 levels increased later, around 9 months of age.

Modification Details

This is a cross between APP23 mice, which overexpress APP751 with the Swedish mutation driven by the murine Thy1 promoter, and PSEN1 knock-in mice expressing human PSEN1 with the R278I mutation under the endogenous presenilin-1 promoter.


For availability information, contact Takaomi Saido or Takashi Saito.

Phenotype Characterization

When visualized, these models will distributed over a 18 month timeline demarcated at the following intervals: 1mo, 3mo, 6mo, 9mo, 12mo, 15mo, 18mo+.


  • Tangles

No Data

  • Neuronal Loss
  • Synaptic Loss
  • Changes in LTP/LTD


By 6 months of age amyloid plaques accumulate in the cortex and hippocampus. A high percentage of plaques are thioflavin-S –positive cored plaques.


Not observed.

Neuronal Loss

No data.


Astrocytosis in the vicinity of plaques in the hippocampus and cortex by 9 months.

Synaptic Loss

No data.

Changes in LTP/LTD

No data.

Cognitive Impairment

Short-term memory deficits are apparent by 3 to 4 months as measured by the Y maze.


  1. In this important paper, Takashi Saito and colleagues explore the pathogenic role of Aβ43 in Alzheimer’s disease (AD) by generating knock-in mice with a presenilin-1 (PSEN1) gene mutation, R278I, which selectively overproduces Aβ43 (1). Aβ42 has been the major focus of interest in amyloidogenesis in AD, and yet various longer species of Aβ, such as Aβ43, Aβ45, Aβ48, Aβ49, and Aβ50 have all been found in AD brains. Saito et al. demonstrate that the R278I mutation causes loss of γ-secretase activity in a recessive manner. Specifically, it appears to inhibit conversion of Aβ43 to Aβ40 by γ-secretase, leading to an increased ratio of Aβ43:Aβ40 and Aβ42:Aβ40, without altering the level of Aβ42. By cross-breeding heterozygous R278I with APP mice, they go on to show that the R278I mutation leads to accelerated Aβ pathology with an accompanying inflammatory response and cognitive impairment, which precedes plaque formation.

    Aβ43 was found to show higher neural toxicity and to contribute more readily to the formation of the thioflavin T-positive β-sheeted structure than either Aβ40 or Aβ42. They also examined brain sections from patients with sporadic AD and found that Aβ43 accumulated more frequently than Aβ40. Interestingly, homozygous PSEN1 R278I knock-in mice were found to have an embryonic lethal phenotype. This was thought to be due to impaired processing of Notch1, another of the substrates of γ-secretase. It would be of great interest to know whether interaction of PSEN1 with other substrates including Notch1 is in any way affected by the heterozygous R278I mutation.

    The clinical phenotype of the R278I PSEN1 mutation has proved to be equally as fascinating. The mutation was originally identified in two individuals with an atypical AD phenotype, who presented with symptoms of language impairment (2). However, we have subsequently studied individuals from another branch of this family who have had either behavioral or typical amnestic presentations of familial AD (FAD). Marked heterogeneity may be witnessed in the clinical phenotype of FAD, between different mutations and even within single families affected by the same mutation (3). Investigating the mechanisms underlying this heterogeneity will be an important direction for further work in the field. As Saito et al. point out, although the majority of PSEN1 mutations are associated with an increased ratio of Aβ42:Aβ40, they have varying effects on the absolute levels of Aβ40 and Aβ42, and future studies should also consider how Aβ43 levels are affected. Investigation of the additional genetic and epigenetic factors that may modify these molecular mechanisms, or their pathological consequences, in different individuals will be particularly challenging. Technological developments including the generation of stem cells from skin fibroblasts of patients with these mutations may provide new opportunities for exploring these issues. In collaboration with John Hardy and Selina Wray, fibroblast cell lines from one of our patients with the R278I PSEN1 mutation have recently been generated. These will be deposited in the Coriell cell repository so that any research group with an interest in the mutation may benefit from open access to this valuable resource. The up-to-date list of lines available upon request from the NINDS repository can be found here.

    Saito et al. conclude their paper by proposing two future lines of investigation into the potential clinical implications of their findings. Firstly, they suggest that measurement of cerebrospinal fluid Aβ43 levels may have value as a potential biomarker for presymptomatic AD. Secondly, they hypothesize that inhibition of Aβ43 production by facilitating conversion of Aβ43 to Aβ40 by the γ-secretase complex may prevent Aβ amyloidosis. With international collaborative initiatives like the Dominantly Inherited Alzheimer Network (DIAN) now studying presymptomatic biomarker changes in FAD mutation carriers and contemplating the design of prevention trials for these individuals (4), interest in such avenues of research could not be more timely.


    . Random mutagenesis of presenilin-1 identifies novel mutants exclusively generating long amyloid beta-peptides. J Biol Chem. 2005 May 13;280(19):19070-7. PubMed.

    . A presenilin 1 R278I mutation presenting with language impairment. Neurology. 2004 Nov 9;63(9):1702-4. PubMed.

    . Correlating familial Alzheimer's disease gene mutations with clinical phenotype. Biomark Med. 2010 Feb;4(1):99-112. PubMed.

    . Autosomal-dominant Alzheimer's disease: a review and proposal for the prevention of Alzheimer's disease. Alzheimers Res Ther. 2011;3(1):1. PubMed.

    View all comments by Martin Rossor
  2. In this interesting and important paper, Saito and colleagues have engineered knock-in mice that cause overproduction of Aβ43. They showed that Aβ43, an overlooked species, was potently amyloidogenic, neurotoxic, and abundant in vivo. Aβ43 may be a new biomarker for AD.

    But I wonder about some of the data. Aβ43 production in R278I/R278I MEF cells or in R278I/+ MEF cells was markedly reduced in a gene dose-dependent manner as shown in supplementary Fig 10bc, suggesting that γ-secretase complex including PS1-R278I clearly impaired Aβ43 production activity as well as Aβ38, Aβ40, and Aβ42. This unambiguously indicates that the R278 mutation is a loss of γ-secretase function. In this figure, an increase in Aβ43 production via the inhibition of Aβ43-to-40 conversion process cannot be observed. As shown in Figure 2k-n, however, a large amount of Aβ43 was detected in conditioned medium from R278I/R278I MEF cells or from R278I/- MEF cells in the ELISA system. Therefore, Aβ43 in conditioned medium may be produced via a PS1-independent processing pathway, not via the inhibition of the Aβ43-to-40 conversion process in the γ-secretase complex.


    . Distinct presenilin-dependent and presenilin-independent gamma-secretases are responsible for total cellular Abeta production. J Neurosci Res. 2003 Nov 1;74(3):361-9. PubMed.

    . A presenilin-independent aspartyl protease prefers the gamma-42 site cleavage. J Neurochem. 2006 Jan;96(1):118-25. PubMed.

    . Pharmacological evidences for DFK167-sensitive presenilin-independent gamma-secretase-like activity. J Neurochem. 2009 Jul;110(1):275-83. PubMed.

    View all comments by Fuyuki Kametani

Make a comment or submit a question

To make a comment you must login or register.


Research Models Citations

  1. APP23

Paper Citations

  1. . Random mutagenesis of presenilin-1 identifies novel mutants exclusively generating long amyloid beta-peptides. J Biol Chem. 2005 May 13;280(19):19070-7. PubMed.
  2. . Potent amyloidogenicity and pathogenicity of Aβ43. Nat Neurosci. 2011 Aug;14(8):1023-32. PubMed.

Other Citations

  1. Takaomi Saido

Further Reading