Another round of cat-and-mouse between scientists and the surreptitious cause of Alzheimer's pathogenesis shows intracellular Aβ lesions lead to age-dependent neurodegeneration and motor impairment in rodents.

Matthias Staufenbiel of the Novartis Institutes for Biomedical Research in Basel, Switzerland, and colleagues created genetically modified mice that produce abundant levels of Aβ40 and/or Aβ42 within their neurons. The location of the transgenic peptides differentiates the work from previous studies, which described mice that secrete the partial APP peptides into the neuronal milieu (McGowan et al., 2005). Although the previously reported mice develop Aβ deposits, they do not show signs of overt toxicity, and so speak to a possibility that intraneuronal Aβ aggregates are among the saboteurs in AD. The new study attempts to explore this possibility.

The study injects new data into an ongoing scientific debate about the role of intraneuronal Aβ, which has recently moved to a focus on the technical challenges of distinguishing Aβ species from their parental precursor, APP (see ARF Webinar).

By linking the partial APP peptides to an N-terminal signal sequence, the researchers forced production of Aβ40 and Aβ42 within the endoplasmic reticulum. The molecular engineering places the peptides in arguably a more natural location within the cell—inside the lumen of the ER and other components of the intracellular membrane system. However, as with other Aβ transgenic mouse studies, the consequences of removing the Aβ peptides from APP are unknown and potentially far reaching. "When you take Aβ out of the context of APP processing, I don't think you really know how relevant the effects are for the disease," said Lars Nilsson of Oslo University in Norway. "The digestion and release of Aβ will change dramatically both in terms of subcellular location and mechanism, which could affect the aggregation process," he said.

Even so, the novel approach to creating Aβ transgenic models can teach us about the basic biology of these proteins, said Nilsson. The authors used histochemical analysis to examine brain sections of mice that expressed either Aβ40 or Aβ42. The latter strain showed significant aggregates and functional problems, while the Aβ40 strain did not. The combined transgenic exhibited intermediate levels of aggregation and pathologies, supporting the possibility that Aβ40 destabilizes the aggregation of Aβ42.

The intraneuronal aggregates were not plaques. Instead, the authors saw Aβ-positive threads within dendrites, dot-like granules in nerve cell bodies, and grain-like structures throughout the gray matter of the brain. Although these strains do not replicate the pathology of AD, the Aβ42 strain does have neurodegeneration and white matter atrophy within the hippocampus and other regions, as well as motor defects, the authors report. The functional impairments evolve over time as the mice age.

While the paper directly demonstrates that Aβ42 aggregates inside neurons cause neuronal death even in young mice, some other scientists express skepticism that the study settles the role of intraneuronal Aβ in AD.

"Direct overexpression of intracellular Aβ1-40 and Aβ1-42 in mice is extremely artificial," wrote Virginia Lee of the University of Pennsylvania School of Medicine in Philadelphia in an e-mail to Alzforum. In AD, Aβ peptides are generated from the proteolytic processing of APP and the majority of the cut materials are secreted—quite a different molecular journey than found in the new transgenic mice.

There is no animal model that recapitulates all or even most of the features of AD. But animal models do allow researchers to test connections between molecular and cellular features of APP and its Aβ offshoots on the one hand, and downstream effects such as white matter loss on the other.

"You can sometimes learn a lot from simply playing around with things and seeing what happens," said Nilsson.—Susan Young

Susan Young is a science writer in Washington, D.C.


  1. As highlighted in the news story, Abramowski and colleagues chose a risky approach in generating mice directly expressing Aβ40 or 42, thereby likely precluding the normal trafficking and localization of Aβ. While we have shied away from studying Aβ biology with such models, reports using different Aβ constructs are nevertheless providing some interesting results. In a study published in Neuron in 2005, McGowan et al. also used artificial constructs, but in their case to secrete Aβ. Interestingly, both the McGowan and Abramowski studies support that Aβ40 can be protective, since coexpression of either the secreted or intracellular Aβ40 with the corresponding Aβ42 constructs reduced brain Aβ42 levels. The present study also underscores how toxic intraneuronal Aβ42 can be compared to the just two-amino-acid-shorter Aβ40. Moreover, in contrast to rising Aβ42 levels in Aβ42 overexpression APP48 mice, Aβ40 levels did not rise in brains of the Aβ40 overexpressing mouse. Intriguingly, pan-neuron expression of Aβ42, but not Aβ40, led to neurodegeneration, particularly of hippocampal pyramidal neurons.

    The current paper describes seeing no extracellular plaques in their intracellular Aβ42 mice, whereas McGowan et al. showed many plaques. Abramowski et al. also note that the Aβ42-secreting mice of McGowan et al. had no obvious behavioral decline or neurodegeneration, while their APP48 mice had both neurodegeneration and functional impairment, albeit motor impairment (given hippocampal cell loss, it would be interesting to look at cognition and memory). From these results, one could conclude that intraneuronal Aβ42 may be toxic, but doesn’t cause typical AD pathology. Yet, looking carefully at Fig. 6H, there clearly is labeling of Aβ42 that does not overlap with MAP2, and even looks plaque-like (MAP2 typically does also label neuron soma). Importantly, a different transgenic mouse model expressing pyro-glut Aβ3-42 was shown to develop diffuse plaques next to marked intraneuronal labeling (Wirths et al., 2009). Interestingly, Abramowski and colleagues describe an age-related Aβ42 increase in neuritic threads, but a decrease in lysosomal granules in APP48 mice. These results parallel what has been described in human Down's syndrome and AD brains, as well as in APP mutant transgenic mice.

    Despite the artificial nature of their constructs, the different models reproduce various aspects of AD, often complementary between studies. The complementarity suggests that both intra- and extracellular Aβ play a role in AD pathogenesis.


    . Intraneuronal pyroglutamate-Abeta 3-42 triggers neurodegeneration and lethal neurological deficits in a transgenic mouse model. Acta Neuropathol. 2009 Oct;118(4):487-96. PubMed.

Make a Comment

To make a comment you must login or register.


Webinar Citations

  1. Intraneuronal Aβ: Was It APP All Along?

Paper Citations

  1. . Abeta42 is essential for parenchymal and vascular amyloid deposition in mice. Neuron. 2005 Jul 21;47(2):191-199. PubMed.

Further Reading

Primary Papers

  1. . Transgenic expression of intraneuronal Aβ42 but not Aβ40 leads to cellular Aβ lesions, degeneration, and functional impairment without typical Alzheimer's disease pathology. J Neurosci. 2012 Jan 25;32(4):1273-83. PubMed.