How exactly do amyloid plaques affect surrounding brain tissue, and does this change over time? Scientists led by Jörg Hanrieder at the University of Gothenburg, Sweden, tackled this question by using isotopically labeled Aβ to timestamp plaques as they formed in mice. As described in an October 11 preprint on bioRxiv, the authors correlated the age of each plaque with its structural characteristics, and examined the effect it had on nearby gene expression using spatial transcriptomics. They found that as plaques matured, they became more compact and fibrillar, and more synaptotoxic.

  • In mice, the oldest amyloid plaques wreaked the greatest synaptic damage.
  • The older the plaques, the more immune response they provoked.
  • These more fibrillar plaques are the most synaptotoxic.

Sean Bendall at Stanford University, California, said these are the first data to confirm the idea that older, more fibrillar plaques cause more havoc. “This is an important contribution that directly links the age of aggregated protein with overall disease pathology—reinforcing that these aggregates are relevant targets for intervention,” he wrote to Alzforum (comment below).

Maturing Plaques. The oldest plaques (left) contain β-sheets (red) and dense fibrillar cores (blue), while younger plaques (middle) lack dense cores, and the youngest (right) lack either. [Courtesy of Wood et al., bioRxiv.]

Scientists still debate how much different types of plaques harm the brain, with some recent evidence suggesting that microglia pack Aβ into deposits to corral the peptide and mitigate the damage (Dec 2009 news; May 2016 news; Apr 2021 news). On the other hand, both lecanemab and donanemab, which target different epitopes on aggregated Aβ, have shown that plaques need to be completely removed for cognitive decline to slow, belying the idea that plaques are neuroprotective (Nov 2023 conference news).

To gain more insight, Hanrieder and colleagues scrutinized what happens in the brain as plaques form and mature. Previously, the scientists had combined mass spectrometry imaging with stable isotope labeling kinetics (iSILK) to track plaque growth in mice. Mice ate chow containing Nitrogen15 for four weeks, allowing this isotope to be incorporated into newly made Aβ. As labeled Aβ deposited into plaques, it in effect timestamped these structures. Later, the scientists took slices from the brains and analyzed plaque proteins via matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI-IMS), providing a molecular fingerprint of differently aged deposits (Jun 2021 news).

In the new work, the scientists combined iSILK with spatial transcriptomics to examine molecular changes around amyloid deposits. Joint first authors Jack Wood and Maciej Dulewicz used APPNL-F knock-in mice, because in these animals, plaques form gradually with age, mimicking what happens in human brain. In these mice, the first sparse plaques appear around 6 months, but amyloidosis does not take off until 9 months. Wood and Maciej fed these mice Nitrogen15 from 6 to 10 months of age, labeling the earliest plaques. At 10 months, the authors sacrificed the mice and examined thin brain slices. When they found plaques that were large enough to extend over two contiguous slices, they analyzed one slice by MALDI-IMS to peg plaque age, and the other for whole-transcriptome-wide expression using NanoString hybridization probes.

The data revealed that the older the plaque, the more synaptic gene expression, particularly cholinergic, dropped in surrounding brain tissue. This jibes with the loss of basal forebrain cholinergic neurons in AD, and with a recent mouse study that found cholinergic synapses wither in the vicinity of plaques (Mufson et al., 1989; Lee and Chen, 2024). Older plaques also provoked higher expression of immune genes, indicating an inflammatory response. The authors did not link gene expression to particular cell types.

When the authors examined 18-month-old mice, the picture changed. The oldest plaques correlated with a boost in metabolic, rather than immune, genes. That suggested to the authors that inflammation was a transient response to initial plaque deposition. The metabolic changes might indicate mitochondrial dysfunction in surrounding cells, they speculated. As in younger mice, the older the plaque, the more it suppressed synaptic and neurotransmitter genes.

Were the responses evoked by younger and older plaques due to physical differences? Using amyloid dyes and fluorescent microscopy, the authors found that the youngest plaques were the least fibrillar, lacking β-sheets. Middle-aged plaques had β-sheets, but still had a loose structure, while the oldest plaques were the most compact and fibrillar (image above). These three types likely correspond to diffuse, fibrillar, and dense-core plaques, the authors noted. Dense-core plaques were associated with the greatest synaptic suppression and surrounded by the most dystrophic neurites, indicating they were the most toxic.

Katherine Sadleir at Northwestern University, Chicago, thinks other explanations are possible. Perhaps the earliest plaques are neurotoxic, but synapses are lost gradually, so that the oldest plaques have the fewest in their vicinity. “This interpretation is supported by microglial inflammatory signatures being most noticeable around the younger plaques and fading with plaque age, replaced by metabolic changes,” she wrote to Alzforum.

How do the data fit with the idea that microglia wall off plaques and prevent them from harming the brain? Hanrieder noted that microglia may be able to contain smaller plaques, but not the larger, dense-core plaques. He was unable to study the effects of smaller plaques because they did not extend across two adjacent tissue slices. In future work, he plans to  determine if smaller plaques are also toxic.—Madolyn Bowman Rogers

Comments

  1. This is a nice use of the pulse/chase proteomic labeling strategy in animal models. Using these isotopic tracing strategies gives us a definitive measurement of the age of structures/proteopathies that are related to neurodegeneration, in a way that cannot be captured otherwise.

    I think the key finding confirms something that has been assumed but never confirmed, which is that longer-lived proteopathies/plaques result in greater molecular pathology. The caveats here are that the mouse model plaques are still much shorter-lived in nature than one would expect accumulation to be in natural aging in the human, as well as the pathology is only inferred from gene expression. Still, this seems to be an important contribution that directly links the age of accumulated pathological proteins with overall disease pathology—reinforcing that these aggregates are relevant targets for intervention.

  2. Methodologically this preprint is novel, using N15 metabolic labeling and mass spectrometry imaging to determine the relative age of individual plaques along with hyperspectral confocal microscopy to characterize the fibrillar state of the amyloid, in addition to spatial transcriptomics to address how plaques of different structures and ages impact the surrounding brain tissue.

    The observations about how plaque fibrillar structure changes as the plaque matures are interesting and support the idea that plaque cores grow more compact over time.  Being able to classify plaques according to their relative age will help further the understanding of how plaques form and grow. Perhaps developing a staging process in mouse models will allow classification of human plaques by age, which could then be combined with spatial transcriptomics to understand the sequence of changes occurring around the plaque over time.

    The authors conclude the older plaques are more neurotoxic based on the fact that levels of synaptic transcripts are more reduced around plaques from 18-month-old mice than from 10-month-old mice. It could also be interpreted that early plaques are neurotoxic and, over time, synapses in the vicinity pf plaques are progressively lost, so there are less synapses the older the plaque is.  This interpretation is supported by the fact that the microglial inflammatory signature is most noticeable around the younger plaques and fades with plaque age, replaced by metabolic changes—perhaps in the microglia and astrocytes.  

    This will be an interesting technology and story to follow as this is published and follow-up studies are done, in other models or in human tissue.

  3. This is an interesting study, applying advanced spatial transcriptomic and isotopic labeling techniques to see how plaque age and fibrillization correlates with adverse local consequences, like loss of synaptic expression. This is an extension of the well-known observation that dense-cored plaques, which take up thioflavin dyes avidly, are associated with neuritic changes and tau aggregation. Diffuse plaques, which take up thioflavin dyes poorly, are less associated with downstream pathologies.

    This data conflicts with a major current conception of drug targets in AD. Lecanemab is purported to target "protofibrils," which are supposedly not fibrils. Toxic protofibrils are thought to lack the well-defined stacks of β-pleated sheet-rich planar monomers held together by inter-molecular backbone hydrogen bonding and intramolecular hydrophobic side-chain interactions (otherwise we would just call them fibrils). Several anti-amyloid antibodies in development also purport to target non-fibrillar oligomers or protofibrils. To my knowledge there has never been a structural definition of these species from human AD brain, and their supposed toxicity is in contrast to the here-repeated observation that the fibrillar plaques (as defined by dyes) are the toxic ones.

    We have found that aqueous preparations from human AD brain, thought to contain "protofibrils," are rich in true amyloid fibrils with the same cryoEM structures as fibrils in plaques, and these can be immunolabeled by lecanemab. In my opinion, true fibrils (rigorously defined as above) at the periphery of dense-cored neuritic plaques are the likeliest target for anti-Aβ therapies, not a structurally different, non-fibrillar protofibril/oligomer. I look forward to presenting at CTAD next week some evidence that lecanemab, aducanumab, and donanemab are all recognizing the same fibrillar species in human AD brain, albeit at different epitopes.

    References:

    . Abundant Aβ fibrils in ultracentrifugal supernatants of aqueous extracts from Alzheimer's disease brains. Neuron. 2023 Jul 5;111(13):2012-2020.e4. Epub 2023 May 10 PubMed.

  4. This pioneering study by Hanrieder and co-workers aims to understand the initiation and progression of Aβ aggregate accumulation and correlations with CNS cell responses by spatial transcriptomics in the proximity of early and late Aβ aggregates. Chemical time stamps were induced using pulse chase of 15N-isotope labeling (iSILK) and mass imaging. As a complementary time stamp, the maturity of the Aβ-plaque structures was monitored by LCO technology developed by us in Linkoping.

    Spatial transcriptomics together with iSILK and LCO/immunofluorescence structure correlations allowed an unprecedented correlation of proximal cell response to the plaque maturity stage. The study concludes that mature Aβ plaques, in comparison with early plaques, regardless of chronological mouse age, were associated with changes suggesting synaptic toxicity responses. This methodological approach puts within reach more information on differentiated cellular responses to plaque fibril structure at different plaque development stages and cell types (microglia, astrocytes, and diverse neuronal populations).

    The choice of the APPNL-F knock-in mouse in this study was clever, because this mouse makes almost exclusively Aβ42, allowing the authors to specifically iSILK-monitor Aβ1-42 species using MALDI-ToF imaging. However, the differentiated LCO signal (qFTAA and hFTAA) for Aβ-plaques in this mouse model is not so strong. The method does differentiate compact fibrillar cores compared to peripheral immature structures in some plaque in APPNL-F as shown by us (Parvin et al., 2024) and in the paper, but for other mouse models it is much more profound.

    We recently showed that the APPNL-F mice displayed very subtle plaque core fibril maturation by LCO hyperspectral microscopy over a rather wide mouse age range, 9-21 months (Parvin et al., 2024). In comparison, transgenic mice such as APP23, which overexpresses Aβ and makes more Aβ40, spectrally overlap with APPNL-F at 9 months, but showed much higher maturation signal of the plaque cores at ages >18 months. The fibrillar maturity of the plaque cores in APP23 was promoted by incorporation of Aβ40 rather than Aβ42 (Parvin et al., 2024). It would be of interest to follow up how the diverse plaque size, Aβ isoform composition, and fibril structure is correlated with cell specific response in different mouse models.

    References:

    . Divergent Age-Dependent Conformational Rearrangement within Aβ Amyloid Deposits in APP23, APPPS1, and AppNL-F Mice. ACS Chem Neurosci. 2024 May 15;15(10):2058-2069. Epub 2024 Apr 23 PubMed.

  5. This study investigates Aβ plaque pathology through a new mass spectrometry imaging method combined with isotope labeling (iSILK) in an Aβ knock-in mouse model (AppNL-F), and couples the plaque age with spatial transcriptomics. This approach enables the study of Aβ plaque formation, maturation, and its related effects on synaptic loss and toxicity across various plaque ages. By enabling more dynamic tracking of plaque maturation than previously achievable, this study addresses a critical gap; traditionally, plaque development has been inferred primarily through postmortem cross-sectional tissue analysis, as done previously in Boon et al., 2020

    The authors present a sophisticated method to map plaque formation and age-related toxicity in an Aβ knock-in mouse model, confirming our observations of increased amyloid fibrillation as plaques mature (Röhr et al., 2020; Boon et al., 2020). However, challenges persist in directly translating these findings to human pathology due to model constraints and limitations in continuous plaque tracking.

    As researchers focused on plaque pathology, we see value in defining the morphologic stages of plaque development for direct comparison with human AD pathology. Human tissue reveals a diversity of Aβ morphologies that most likely does not align with a single maturation trajectory (Thal et al., 2006; Capetillo-Zarate et al. 2006Boon et al. 2020). 

    Extending the analysis to include specific plaque types such as coarse-grained plaques, cotton wool plaques, and cerebral amyloid angiopathy will be particularly insightful in future studies.

    References:

    . The coarse-grained plaque: a divergent Aβ plaque-type in early-onset Alzheimer's disease. Acta Neuropathol. 2020 Dec;140(6):811-830. Epub 2020 Sep 14 PubMed.

    . Label-free vibrational imaging of different Aβ plaque types in Alzheimer's disease reveals sequential events in plaque development. Acta Neuropathol Commun. 2020 Dec 11;8(1):222. PubMed.

    . The development of amyloid beta protein deposits in the aged brain. Sci Aging Knowledge Environ. 2006 Mar 8;2006(6):re1. PubMed.

    . Selective vulnerability of different types of commissural neurons for amyloid beta-protein-induced neurodegeneration in APP23 mice correlates with dendritic tree morphology. Brain. 2006 Nov;129(Pt 11):2992-3005. PubMed.

  6. This study elegantly places a timetable on plaque formation in a transgenic mouse model (AppNL-F) selected to develop amyloid pathology relatively slowly, using stable isotope labeling, mass spectrometry, advanced microscopy and spatial transcriptomic analyses. One insight is that small dense cored deposits occur early, followed by more widespread "diffuse" deposition. The methods confirm that amyloid deposition is more abundant and early in the cortex than the hippocampus.

    Several of these general findings are known from a variety of prior studies on plaques. However, this study provides timestamping of the amyloid, and the spatial transcriptomic methods enabled the identification of early downregulation of synapse-related genes and upregulation of inflammation pathways, and at later timepoints, an increase in metabolism-related genes. The finding that immune activation and synaptic downregulation co-occur does not inform whether there is a mechanistic relationship between these, although many possibilities exist.

    This combination of elegant approaches will be interesting to apply to other pathogenic players in AD, including tau pathogenesis, and also could provide insights into what happens when the mouse is treated with immunotherapy that can clear the amyloid.

  7. This study advances our understanding of amyloid plaque formation and maturation in AD by building on previous isotope labeling and mass spectrometry imaging techniques, while also introducing several key innovations. It utilizes the AppNL-F mouse model, which more gradually develops amyloid plaques, closely mimicking the slow progression observed in human AD, than the more aggressive AppNL-G-F model used previously. This choice of model, coupled with advanced techniques—including iSILK (isotope-encoded stable isotope labeling kinetics) mass spectrometry imaging, hyperspectral microscopy, and spatial transcriptomics—allows for a more comprehensive view of amyloid pathology as it progresses.

    In addition to its technical advancements, this study examines plaques at both early (10 months) and later (18 months) stages, providing a comparative view of disease progression. Prior research has often focused on static features of plaques, but this study highlights their dynamic changes over time, revealing that plaque age correlates with structural maturity and, importantly, with increased toxicity. The findings demonstrate that as plaques mature, they become more compact and contribute to greater synapse loss and increased neuronal damage in their surrounding microenvironment. Furthermore, the study identifies three distinct categories of plaque, each with unique structural properties and impact, highlighting a nuanced view of amyloid pathology’s progression in AD.

    The methods for tracking plaque maturation are noteworthy. iSILK allows for precise temporal tracking by incorporating 15N isotopes into newly synthesized proteins, effectively distinguishing between “old” and “new” amyloid deposits based on their isotopic signature. This technique provides an accurate picture of plaque formation and aging by tracking deposits based on their actual maturation process, rather than merely their chronological age. By leveraging this innovation, the study presents an accurate and dynamic framework for understanding how individual plaques evolve and become more toxic.

    Overall, this study adds a layer of specificity to AD research, suggesting that the age and structural maturity of plaques—not just their presence—are key factors in their neurotoxic effects. This refined understanding of plaque evolution and its impact on surrounding tissue highlights potential new targets for therapeutic strategies, with an emphasis on early intervention before plaques mature into highly compact, damaging structures.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Long Life With Tight Plaques—Repressing IGF-1 Protects AD Mice
  2. Barrier Function: TREM2 Helps Microglia to Compact Amyloid Plaques
  3. Microglia Build Plaques to Protect the Brain
  4. Gotta Get Rid of It All: Total Plaque Clearance Key for Clinical Benefit
  5. iSILK Tracks Growth of Mouse Plaques at Peptide Level

Therapeutics Citations

  1. Leqembi
  2. Donanemab

Research Models Citations

  1. APP NL-F Knock-in

Paper Citations

  1. . Loss of nerve growth factor receptor-containing neurons in Alzheimer's disease: a quantitative analysis across subregions of the basal forebrain. Exp Neurol. 1989 Sep;105(3):221-32. PubMed.
  2. . Loss of Cholinergic and Monoaminergic Afferents in APPswe/PS1ΔE9 Transgenic Mouse Model of Cerebral Amyloidosis Preferentially Occurs Near Amyloid Plaques. Int J Mol Sci. 2024 May 3;25(9) PubMed.

Further Reading

Primary Papers

  1. . Isotope Encoded Chemical Imaging Identifies Amyloid Plaque Age Dependent Structural Maturation, Synaptic Loss, and Increased Toxicity. 2024 Oct 11 10.1101/2024.10.08.617019 (version 1) bioRxiv.