As Alzheimer’s disease develops, formerly innocuous Aβ peptides turn rogue, assuming toxic forms and damaging synaptic functions. In the March 13 Nature Communications, researchers led by Gunnar Gouras at Lund University, Sweden, now provide a glimpse of the earliest stages of this process in a mouse brain. They used Fourier transform infrared microspectroscopy (μFTIR) to spot the initial appearance of small Aβ aggregates that have a β-sheet structure. These accumulated in cortical postsynapses about a month before plaques became visible by conventional staining. “The data suggest that plaques don’t just pop up overnight, as previous studies suggest. There are early ongoing changes that are invisible to PET and multiphoton microscopy,” Gouras noted.

Others agreed that μFTIR could become a valuable technique for examining subtle structural changes in Aβ and other aggregating proteins. “This works confirms our concepts of what is happening in the brain over time as it progresses from normal to pathologic histology. Thanks to μFTIR, we can see it happen,” said David Teplow at the University of California, Los Angeles.

β-Sheets Form Early. The Aβ β-sheet signal (yellow and red) is absent in the transgenic APP mouse cortex at one month (left), appears at two (middle), and intensifies at three (right), when plaques develop. [Courtesy of Klementieva et al., Nature Communications.]

Plaques are made up of amyloid arranged in β-sheets, and can pop up within a day in repeat multiphoton microscopy images of the mouse brain (see Feb 2008 news). Aβ oligomers are now believed to be the more toxic form of the peptide, however, and because they are typically invisible to microscopy, exactly when and where they appear has remained mysterious.

To shed light on this, the authors turned to μFTIR, which provides information about the chemical structure of molecules based on the wavelengths of infrared radiation they absorb. The technique can distinguish between native β-sheet structure within proteins versus linked β-sheets in amyloid fibrils, and can be used to visualize protein aggregation in cells and tissues (see Zandomeneghi et al., 2004; Miller et al., 2013). 

First author Oxana Klementieva prepared freeze-dried sections from the cortices of one-, two-, and three-month-old Tg19959 mice. Developed by Gouras’ group, this model carries human APP harboring the Swedish and Indiana mutations, and develops amyloid plaques at three months of age (see Li et al., 2004). A μFTIR signal indicative of Aβ β-sheets appeared at two months, and intensified at three, when plaques became visible (see image above). “We were surprised at the amount of β-sheet structure already present at two months,” Gouras said. The nature of these species remains unclear but they appear to be much larger than monomers, but not fibrillar.

To pinpoint these Aβ aggregates, the authors labeled cortical sections with antibodies against the Aβ42 C-terminus and the presynaptic marker synaptophysin and postsynaptic marker drebrin, and analyzed them with three-dimensional confocal microscopy. They saw Aβ42 piling up at postsynapses at two months, although the antibody could not distinguish whether this was monomeric or oligomeric, or what secondary structure it had. At the same time, postsynapses began to intertwine with presynapses, indicating gross structural defects at the synapse. By three months, neurites appeared dystrophic (see image below). “This localization is really interesting, because the synapse is a hotspot in Alzheimer’s disease,” Teplow noted.

Synaptic Carnage. As Aβ42 (red) increases in two-month-old transgenics (middle), presynapses (purple), and postsynapses (green) begin to merge (overlay appears white) compared to wild-type synapses (left); by three months (right), neurites have swollen and look dystrophic. [Courtesy of Klementieva et al., Nature Communications.]

The authors also homogenized brain samples from the transgenics and ran the proteins on non-denaturing blue native gels to get a better idea of the size of Aβ aggregates. Intriguingly, Aβ42 from wild-types and one-month-old transgenics ran at the same speed as a synthetic Aβ42 tetramer. This band became less abundant in two-month-old transgenics, as higher molecular weight Aβ aggregates took over. The data hint that physiological Aβ42 may exist as a tetramer, and only when this falls apart does the peptide assume toxic forms, Gouras suggested. If confirmed, he believes this could have therapeutic implications. The peripheral amyloidosis familial amyloid polyneuropathy (FAP) has been successfully treated by stabilizing the amyloidogenic protein, transthyretin, in its native tetrameric form, and some research indicates native α-synuclein exists as a tetramer as well (see Aug 2011 news; Dec 2013 news; Aug 2011 news). 

Lawrence Rajendran at the University of Zurich found the ideas in the paper intriguing. “By elegantly combining biophysics and in vivo structural biology, the authors identify the possible existence of tetrameric Aβ as well as a conformational change before plaque deposition occurs,” he wrote to Alzforum. “We might need to be looking at tetramers as a pre-pathological form,” he added.

However, other researchers said more biochemical analysis will be needed to confirm the existence of an in vivo tetramer. “Previous studies have shown that different forms of Aβ can have an aberrant mobility in native and denaturing gels,” Hilal Lashuel at the Swiss Federal Institute of Technology (EPFL), Lausanne, wrote to Alzforum. Teplow pointed out that other proteins could be bound to Aβ42, causing it to run like a heavier molecule in the gel. Others also urged caution regarding the idea of stabilizing tetramers, noting that several previous studies have found tetrameric Aβ to be toxic in vitro (see Jun 2009 news). 

Gouras believes that the putative Aβ42 tetramers might incorporate lipids or other modifications that make them distinct from synthetic Aβ42 tetramers. The structures isolated from mice react with an antibody that normally recognizes only monomeric Aβ and does not recognize synthetic Aβ42 tetramers, suggesting there is something unique about their structure, Gouras said. He plans to characterize them further.—Madolyn Bowman Rogers


  1. By elegantly combining biophysics and in vivo structural biology, the authors identify the possible existence of tetrameric Aβ and the conformational change before plaque deposition occurs. This could suggest that, like transthyretin and the infamous α-synuclein, we might need to be looking at tetramers as a “prepathological form,” and we might need to brace ourselves for the fact that stabilizing tetrameric Aβ could be a possibility for therapy. Aβ-APP binding and a possible influence on the subsequent APP processing are also potential avenues for therapy. Gunnar Gouras is arguably one of the fine minds in the AD field, and his latest work gives new insights and makes us rethink our therapeutic deliberations for AD.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Popcorn Plaque? Alzheimer Disease Is Slow, Yet Plaque Growth Is Fast
  2. Amyloid-Blocking Drug Poised for Approval for Rare Disease
  3. Artificial Chaperone Keeps Amyloid-Forming Protein in Check
  4. An α-Synuclein Twist—Native Protein a Helical Tetramer
  5. The Toxic Fold? Aβ Dodecamers, Tetramers Show Their Conformations

Paper Citations

  1. . FTIR reveals structural differences between native beta-sheet proteins and amyloid fibrils. Protein Sci. 2004 Dec;13(12):3314-21. Epub 2004 Nov 10 PubMed.
  2. . FTIR spectroscopic imaging of protein aggregation in living cells. Biochim Biophys Acta. 2013 Oct;1828(10):2339-46. Epub 2013 Jan 25 PubMed.
  3. . Increased plaque burden in brains of APP mutant MnSOD heterozygous knockout mice. J Neurochem. 2004 Jun;89(5):1308-12. PubMed.

Further Reading

Primary Papers

  1. . Pre-plaque conformational changes in Alzheimer's disease-linked Aβ and APP. Nat Commun. 2017 Mar 13;8:14726. PubMed.