The remarkable properties of amyloid-β derived from human Alzheimer’s disease brain: swinging the streetlight

This scientific commentary refers to ‘Highly potent soluble amyloid-β seeds in human Alzheimer brain but not cerebrospinal fluid’ by Fritschi et al. (doi: 10.1093/brain/awu255).

Despite many years of intensive study, the specific amyloid-β species in the human brain responsible for the pathophysiological processes underlying Alzheimer’s disease have yet to be identified. In part, this may be because we have been ‘searching under the streetlight’: examination of other sources of amyloid-β such as synthetic preparations, material derived from the brains of transgenic animals, and amyloid-β recovered from human CSF after lumbar puncture has been much more convenient than direct evaluation of human brain-derived material. However, telltale signs over the last several years have indicated that amyloid-β derived from human brain may have properties or constituents that are qualitatively and quantitatively different from those of amyloid-β from other sources. The paper by Fritschi et al. (2014) from Mathias Jucker’s group in this issue of Brain is a major contribution to this line of investigation. Notably, the paper provides perhaps the most definitive evidence yet that human brain-derived amyloid-β has fundamentally dissimilar properties to human CSF-derived material.

Specifically, Fritschi et al. demonstrate that <1 attomole (10⁻¹⁸ moles) of water-soluble amyloid-β from frozen human Alzheimer’s disease brain tissue is sufficient to induce accelerated amyloid-β plaque deposition when injected into the brains of young APP transgenic mice. In contrast, human CSF from patients with dementia of the Alzheimer’s type containing over 100 000 times greater total mass of amyloid-β failed to affect amyloid-deposition when injected in the same manner. Neither intrinsic anti-aggregation activity in CSF, nor freeze-thaw effects could explain these results; and CSF from transgenic mice behaved similarly, excluding any effect of species. The same lab has previously reported that synthetic amyloid-β at 100–1000-fold higher concentrations also fails to induce plaque deposition (Meyer-Luehmann et al., 2006). Furthermore, neither synthetic amyloid-β dimers nor protofibrils, mixed synthetic amyloid-β plus astrocyte-derived apolipoprotein E particles, nor cell-culture derived amyloid-β immunoreactive species induce in vivo plaque deposition (Meyer-Luehmann et al., 2006).

The explanation for these remarkable observations is not at all clear. By way of tantalizing hints, Fritschi et al. reveal that the amyloid-β immunoreactive particles are larger in the human Alzheimer’s disease brain lysates than in the CSF, and their in vitro particle seeding activity is higher. Furthermore, the human brain material contains readily detectable N-terminally truncated amyloid-β species such as amyloid-β_4–40 and amyloid-β_4–42, whereas CSF contains primarily full length amyloid-β and C-terminally truncated species such as amyloid-β_1–17 and amyloid-β_1–38. The truncation analyses were based on matrix-assisted laser desorption/ionization (MALDI) mass spectrometry. An important future direction will be to include higher resolution assessments of the material using techniques such as liquid chromatography with tandem mass spectrometry (LC-MS/MS) to build confidence in the identifications of amyloid-β species given the complexity of the material. Additional mass spectrometry approaches to identify other post-translational modifications and co-associated proteins will also be of great interest. Furthermore, while the observations of relative differences are likely to be solid and are based on well-designed controls, quantitation of amyloid-β at the attomole level is certainly challenging, particularly given its propensity for non-linear behaviour. Future refinements will be needed in order to perform quantitatively precise studies at this scale.

The results presented by Fritschi et al. are perhaps the most dramatic example of an emerging theme; the vast functional diversity within the ‘amyloid-β-ome’. Other examples include the following: Noguchi et al. (2009) reported that large soluble amyloid-β-containing assemblies termed amylospheroids, extracted from human Alzheimer’s disease brain lysates, caused apoptosis of rat septal neurons in culture at concentrations ∼40-fold lower than required for synthetic amyloid-β aggregates of similar size and immunoreactivity profile. Jin et al. (2011) revealed that small ( ∼ 8 kDa) amyloid-β immunoreactive species derived from human Alzheimer’s disease brain caused cytoskeletal disruption in cultured rat hippocampal neurons at concentrations 1000-fold lower than required for synthetic amyloid-β dimers. Moreover, Langer et al. (2011) found that a sub-fraction accounting for <0.05% of total brain amyloid-β from transgenic mouse brain was responsible for 30% of amyloid-β aggregate seeding activity in vivo (Langer et al., 2011).

An important avenue for future investigations will be to determine the specific relationship between these N-terminal modifications of amyloid-β and induction of amyloid-β aggregation in vivo. The specific forms of amyloid-β detected by mass spectrometry may be just the tip of the iceberg; many other post-translationally modified forms of amyloid-β have been described recently (Bayer and Wirths, 2014). Furthermore, the ultimate biological activity is likely to be governed most directly by the 3D structures of amyloid-β assemblies, which may require advanced methods of assessment such as pulsed hydrogen–deuterium exchange and fast photochemical oxidation mass spectrometry for characterization (Gau et al., 2013; Zhang et al., 2013).

Importantly, the relationship between aggregate seeding activity and synaptic toxicity, tau-related pathophysiological processes, and inflammatory responses remains to be established; it is possible in principle that these events are dissociable with different structural determinants. However, oligomeric forms of amyloid-β derived from human Alzheimer’s disease brain seem to be the most potent triggers for neurotoxicity, and in reports analogous to the paper by Fritschi et al., amyloid-β oligomers seems to be present in human Alzheimer’s disease CSF at concentrations many orders of magnitude lower than in brain tissue lysates, if at all (Xia et al., 2009; Esparza et al., 2013; Savage et al., 2014).

In conclusion, with the efforts of Fritschi et al., as well as several other groups in recent years, the streetlight is now beginning to swing around to illuminate the human brain directly.

Hopefully this is where we will find the keys to developing effective therapeutics for Alzheimer’s disease.

References