Accumulation of organized, self-polymerizing protein aggregates is a hallmark of Alzheimer's disease and infectious prion diseases. The similarities between these conditions may be even closer than that.
Amyloid fibrils are malicious. These insoluble, highly organized protein aggregates are associated with devastating disorders such as Alzheimer's and Parkinson's diseases, type II diabetes and the prion (proteinaceous infectious particle) diseases that include Creutzfeldt–Jakob and mad cow diseases1. The aggregates are toxic, highly stable and can jump-start self-polymerization by recruiting their normal, soluble protein counterpart2. Amyloid self-polymerization is also the basis of the 'protein-only' hypothesis for the mechanism of prion infectivity3: the infectious prion conformation replicates itself in a host by catching the 'benign' host prion protein and forcing it into the infectious conformation. The nature of this two-component system means that the conformations of both the prion agent and the host prion protein influence the characteristics (phenotype) of the resulting prion strains, such as age of onset of the disease in the host or brain localization of the amyloid4.
If a prion amyloid is infectious, can other amyloid disorders such as Alzheimer's disease also be infectious? This is the provocative question raised by Meyer-Luehmann et al.5 in Science. They report data showing that amyloids of the amyloid-β (Aβ) peptide that is associated with Alzheimer's disease behave like an infectious agent when injected into the brain of a mouse model of Alzheimer's disease, generating phenotypes that depend on both the host and the agent.
Aggregation of the Aβ peptide into amyloid fibrils is a central process in Alzheimer's and other diseases1 (Fig. 1). In vitro studies show that amyloid formation is affected by both elapsed time and Aβ concentration, but that it can be rapidly initiated by the addition of a 'seed' consisting of Aβ amyloids2.
Meyer-Luehmann et al.5 now provide in vivo evidence of such a time- and concentration-dependent seeded aggregation of Aβ. Their experiments used brain extracts from deceased Alzheimer's patients or from aged mice genetically predisposed to develop an Alzheimer's-like disease (transgenic Alzheimer's mice), or synthetic Aβ. The extracts were injected into the brains of young, presymptomatic transgenic Alzheimer's mice. In all instances, the authors observed seeding activity (although with reduced efficiency in the case of synthetic amyloid). Aβ peptides seem to be responsible for the seeding activity because this activity was lost if injected extracts were used in which either the Aβ conformation had been destroyed or Aβ removed. These studies further strengthen observations of in vivo seeding of Aβ in transgenic mice6 and in non-human primates7, suggesting a mechanism reminiscent of prion infectivity. Furthermore, the pathology of Aβ formation is governed by both the injected agent and the host, features typically observed for prion strains.
Meyer-Luehmann et al.5 used two distinct transgenic mouse models of Alzheimer's disease: APP23 mice and APPPS1 mice. APP23 mice overproduce mainly the 40-amino-acid Aβ peptide, which results in the deposition of predominantly diffuse and filamentous Aβ in aged animals. APPPS1 mice overproduce mainly the 42-amino-acid Aβ, and show a compact (punctate) deposition of Aβ with age. Injection of brain extracts from aged APPPS1 mice into young APP23 hosts consistently induced punctate Aβ deposition that was mainly confined to one part of the brain (the subgranular layer of the hippocampus). In contrast, extracts from aged APP23 mice injected into APP23 hosts yielded primarily diffuse and filamentous lesions, with substantial diffuse Aβ deposition. When the experiments were applied to the APPPS1 host, deposition induced by the APPPS1 extract was even more punctate, whereas that induced by the APP23 extract was a mixture of the filamentous and punctate Aβ forms.
What factors might make an amyloid infectious? According to a mathematical model8 validated by experiments in three strains of yeast prion, the generation of a prion and its phenotype depends on several factors: the amount of amyloid; the conformation-dependent growth rate of the amyloid; the division rate of the amyloid into seeds; and the concentration of the soluble protein counterpart.
On this account, manipulation of any of these variables can transform any amyloid into a prion. For example, an increase of soluble amyloid protein in the host might be sufficient to transform an amyloid into a prion. In contrast, the formation of long fibrils decreases the ratio of the number of seeds to amyloid protein, reducing the seeding capacity. The amyloid conformation determines its growth and division rates, thereby influencing the titre of infectivity and the prion phenotype. Indeed, Meyer-Luehmann et. al.5 observed induction of Aβ formation by Aβ injection only in Aβ-overproducing transgenic mice and not in normal mice. Injection of long, synthetic Aβ fibrils resulted in a low seeding capacity, and the different amyloid phenotypes described might be attributed to the distinct properties and three-dimensional structures of amyloids consisting of either the 40- or the 42-amino-acid Aβ9,10.
It might sound shocking that Alzheimer's disease, and possibly other amyloid diseases, can be infectious under certain circumstances (in this case5, intracerebral injection into Aβ-overproducing mice). According to the nucleation–polymerization model of amyloids, however, it is to be expected. So where do we go from here?
The experimental evidence5,6,7 and widespread incidence of amyloid diseases highlight the need for basic, clinical and epidemiological studies into the possible transmission of amyloid diseases between humans. In the United States alone, 4.5 million people suffer from Alzheimer's disease, and 20 million from type II diabetes; in particular, we need to find out whether seeds can be transmitted and can accelerate the onset of disease in humans.
For those involved in basic research, there are plenty of other issues to tackle. Is the use of mice that overexpress prion protein still adequate to study transmissible spongiform encephalopathies, given that overexpression might mask the real properties of the agent11? How many strains of Alzheimer's disease are there, and how well are they reflected in the established transgenic mouse strains? If there are several strains — and so several amyloid conformations — how relevant are the structural studies of amyloids that are generated in vitro? Amyloids are certainly malicious in their association with human disease. They are also highly challenging in the increasing number of questions they pose for researchers.
Selkoe, D. J. Nature 426, 900–904 (2003).
Harper, J. D. & Lansbury, P. T. Annu. Rev. Biochem. 66, 385–407 (1997).
Prusiner, S. Proc. Natl Acad. Sci. USA 95, 13363–13383 (1998).
Weissman, C. Nature Rev. Microbiol. 2, 861–871 (2004).
Meyer-Luehmann, M. et al. Science 313, 1781–1784 (2006).
Kane, M. D. et al. J. Neurosci. 20, 3606–3611 (2000).
Ridley, R. M. et al. J. Neural Transm. 113, 1243–1251 (2006).
Tanaka, M. et al. Nature 442, 585–589 (2006).
Petkova, A. T. et al. Biochemistry 45, 498–512 (2006).
Luhrs, T. et al. Proc. Natl Acad. Sci. USA 102, 17342–17347 (2005).
Aguzzi, A. & Heikenwalder, M. Nature Rev. Microbiol. 4, 765–775 (2006).
About this article
Nature Nanotechnology (2011)
Nature Reviews Molecular Cell Biology (2007)