Gilbert MA, Fatima N, Jenkins J, O'Sullivan TJ, Schertel A, Halfon Y, Morrema TH, Geibel M, Radford SE, Hoozemans JJ, Frank RA. In situ cryo-electron tomography of beta-amyloid and tau in post-mortem Alzheimer's disease brain. 2023 Jul 18 10.1101/2023.07.17.549278 (version 1) bioRxiv.

Recommends

Please login to recommend the paper.

Comments

  1. Frank and colleagues now extend their beautiful cryo-ET work in mouse models to postmortem human Alzheimer’s disease brain with largely similar results. This work goes beyond the plastic-embedded, negative-staining EM begun in the last century to show us in three dimensions how Aβ and tau fibrils intertwine with cellular structures, and particularly for tau, details of how the atomic structure of fibrils correlates with their location. The past several years of work from the labs of Michel Goedert and Sjors Scheres, among others, have shown us how the atomic structure of extracted fibrils correlates with histopathologic diagnosis. These new findings from the Frank lab inch us closer to understanding how the atomic structure of amyloid fibrils relates to the adjacent subcellular disturbances underlying dementia.

    It is tempting to conclude that the branching seen in some Aβ fibrils reflects secondary nucleation and that the narrower “protofilament-like rods” are indeed Aβ protofilaments (i.e., one stack of monomers without a pair as seen in all extracted Aβ fibrils to date). This may be the case. However, the resolution is not sufficient to exclude another explanation, such as a non-Aβ molecule interdigitated with and/or bound to the Aβ fibrils. If the fibril bundles and lattices can interdigitate with cellular structures so intimately, it would not be surprising that other molecules could interdigitate with individual fibrils, too. We look forward to the Frank group extending their already groundbreaking technological advancements to test this hypothesis.

    View all comments by Dennis Selkoe

  2. This preprint by Gilbert, Fatima, Jenkins, O'Sullivan et al. from the group of René Frank at the University of Leeds describes the first-ever use of electron cryo-tomography (cryo-ET) on high-pressure frozen brain tissue (temporal cortex) from an individual with Alzheimer’s disease. It opens the way to future cryo-ET studies of diseased human brain tissues.

    The work follows on from a recent study by the same group of cryo-ET of Aβ plaques from the cerebral cortices of AppNL-G-F knock-in mice (Leistner al., 2023). Similar to their previous findings, the authors now report that Aβ plaques from human brain contained abundant filaments that were intermingled with vesicles and droplets. It remains to be seen if Aβ 42 filaments were present other than those extracted from the brains of individuals with Alzheimer’s disease (Yang et al., 2022). Moreover, parallel clusters of unbranched filaments with the morphology of paired helical filaments (PHFs) were observed, indicating that PHFs can be identified in human brain samples by cryo-ET. It also remains to be seen if this is true of filaments with the chronic traumatic encephalopathy (CTE) fold, as suggested by the authors. Most cases of Alzheimer’s disease lack CTE filaments. It would therefore be important to demonstrate their presence in the brain of this individual using a complementary technique, such as cryo-EM of sarkosyl-insoluble tau filaments.

    Even though these findings break new ground, they were not obtained using “fresh” tissue samples, contrary to what is being claimed. Traditionally, one distinguishes between fresh, frozen, and fixed tissues. The findings described here used frozen brain samples. Thus, following a postmortem delay of around six hours, the brain tissues were frozen in liquid nitrogen, which will have led to the formation of crystalline ice and subsequent tissue damage. Before cryo-ET, the tissues were thawed and frozen again under high pressure. To investigate deleterious effects of this process on Aβ plaques, the authors could perform control experiments using AppNL-G-F mice.

    Many years ago, fixed brain tissues from individuals with Alzheimer’s disease were used to identify the abnormal filaments that were named PHFs (Kidd, 1963) and to visualize the ultrastructure of plaques and tangles (Terry et al., 1964). It will be interesting to see if cryo-ET on fresh brain samples from individuals with Alzheimer’s disease can provide fundamentally new insights.  

    It will be difficult to get around problems posed by the changes in brain chemistry that occur postmortem. For example, it was shown previously that non-assembled human tau is rapidly dephosphorylated after death (Matsuo et al., 1994). The future lies probably in the ability to perform cryo-ET on human brain biopsy samples or on fresh tissues obtained after very short postmortem delays.

    References:

    Kidd M. Paired helical filaments in electron microscopy of Alzheimer's disease. Nature. 1963 Jan;197:192-3.

    Leistner C, Wilkinson M, Burgess A, Lovatt M, Goodbody S, Xu Y, Deuchars S, Radford SE, Ranson NA, Frank RA. The in-tissue molecular architecture of β-amyloid pathology in the mammalian brain. Nat Commun. 2023 May 17;14(1):2833. PubMed.

    Matsuo ES, Shin RW, Billingsley ML, Van deVoorde A, O'Connor M, Trojanowski JQ, Lee VM. Biopsy-derived adult human brain tau is phosphorylated at many of the same sites as Alzheimer's disease paired helical filament tau. Neuron. 1994 Oct;13(4):989-1002. PubMed.

    Terry RD, Gonatas NK, Weiss M. Ultrastructural studies in Alzheimer's presenile dementia. Am J Pathol. 1964 Feb;44:269-87.

    Yang Y, Arseni D, Zhang W, Huang M, Lövestam S, Schweighauser M, Kotecha A, Murzin AG, Peak-Chew SY, Macdonald J, Lavenir I, Garringer HJ, Gelpi E, Newell KL, Kovacs GG, Vidal R, Ghetti B, Ryskeldi-Falcon B, Scheres SH, Goedert M. Cryo-EM structures of amyloid-β 42 filaments from human brains. Science. 2022 Jan 14;375(6577):167-172. Epub 2022 Jan 13 PubMed.

    View all comments by Michel Goedert

  3. First authors Madeleine Gilbert, Nayab Fatima, Joshua Jenkins, and Thomas O’Sullivan, from the group of René Frank at the University of Leeds, have pioneered the use of electron cryo-tomography (cryo-ET) to image Aβ and tau pathology at molecular resolution in unfixed frozen human brain tissue.

    This is a technical feat. The authors first used fluorescence cryo-microscopy to locate Aβ and tau pathology in high-pressure frozen tissue stained with the amyloidophilic dye methoxy-X04. This guided cryo-ultramicrotomy to produce ultra-thin tissue sections suitable for cryo-ET. To improve quality for a subset of tomograms, the authors also used focused ion beam cryo-milling to excise thin lamellae from the tissue.

    Using this workflow, the authors imaged a parenchymal Aβ plaque, tau neuropil threads, and an extracellular tau deposit in tissue from the temporal cortex of an individual who had AD. They show in exquisite detail the presence of amyloid filaments and membranous structures, consistent with electron microscopy of Aβ deposits and tau inclusions in fixed human brain tissue (Kidd, 1963; Terry, 1963; Kidd, 1964; Terry et al., 1964). It will be fascinating to see what additional molecular pathology is revealed by the higher-resolution and better tissue preservation of cryo-ET.

    The authors also performed sub-tomogram averaging of filaments in tau inclusions. The resulting reconstructions were of sufficient resolution to identify tau paired helical filaments (PHFs) by their protofilament ultrastructure. Filaments from one inclusion also produced a reconstruction that resembled the ultrastructure of type I CTE tau filaments (Falcon et al., 2019), raising the hypothesis that this individual may have suffered from a neuroinflammatory insult (Qi et al., 2023, 2023). Tau straight filaments (SFs) were not observed. We recently used cryo-ET and sub-tomogram averaging to identify PHFs and SFs tethered within extracellular vesicles from the brains of individuals with AD (Fowler et al., 2023). Implementing a three-dimensional classification step during sub-tomogram averaging enabled us to identify SFs.

    The approach used here for targeted cryo-ET of amyloid pathology in human brain tissue has huge potential, including in the study of fresh tissue, for uncovering additional pathology in AD and in other neurodegenerative diseases.

    References:

    Falcon B, Zivanov J, Zhang W, Murzin AG, Garringer HJ, Vidal R, Crowther RA, Newell KL, Ghetti B, Goedert M, Scheres SH. Novel tau filament fold in chronic traumatic encephalopathy encloses hydrophobic molecules. Nature. 2019 Apr;568(7752):420-423. Epub 2019 Mar 20 PubMed.

    Fowler SL, Behr TS, Turkes E, Cauhy PM, Foiani MS, Schaler A, Crowley G, Bez S, Ficulle E, Tsefou E, O'Brien DP, Fischer R, Geary B, Gaur P, Miller C, D'Acunzo P, Levy E, Duff KE, Ryskeldi-Falcon B. Tau filaments are tethered within brain extracellular vesicles in Alzheimer's disease. bioRxiv. 2023 Apr 30; PubMed.

    Kidd M. Paired helical filaments in electron microscopy of Alzheimer's disease. Nature. 1963 Jan;197:192-3.

    Kidd M. ALZHEIMER'S DISEASE--AN ELECTRON MICROSCOPICAL STUDY. Brain. 1964 Jun;87:307-20. PubMed.

    Qi C, Hasegawa M, Takao M, Sakai M, Sasaki M, Mizutani M, Akagi A, Iwasaki Y, Miyahara H, Yoshida M, Scheres SH, Goedert M. Identical tau filaments in subacute sclerosing panencephalitis and chronic traumatic encephalopathy. Acta Neuropathol Commun. 2023 May 5;11(1):74. PubMed.

    Qi C, Verheijen BM, Kokubo Y, Shi Y, Tetter S, Murzin AG, Nakahara A, Morimoto S, Vermulst M, Sasaki R, Aronica E, Hirokawa Y, Oyanagi K, Kakita A, Ryskeldi-Falcon B, Yoshida M, Hasegawa M, Scheres SH, Goedert M. Tau Filaments from Amyotrophic Lateral Sclerosis/Parkinsonism-Dementia Complex (ALS/PDC) adopt the CTE Fold. bioRxiv. 2023 Apr 28; PubMed.

    TERRY RD. THE FINE STRUCTURE OF NEUROFIBRILLARY TANGLES IN ALZHEIMER'S DISEASE. J Neuropathol Exp Neurol. 1963 Oct;22:629-42. PubMed.

    Terry RD, Gonatas NK, Weiss M. Ultrastructural studies in Alzheimer's presenile dementia. Am J Pathol. 1964 Feb;44:269-87.

    View all comments by Tiana Sophia Behr

  4. This study is exciting because it reproduced in human postmortem AD brain tissue prior findings from mouse models, showing the accumulation of extracellular vesicles (EVs), as well as lipid droplets and different form of amyloid fibrils, in amyloid plaques. The authors also detected open lipid membranes in plaques, indicative of damaged cells, which can release intracellular vesicles.

    APP NL-G-F mice do not develop tau pathology, but this study extends our understanding of the Cryo-EM structure to tau tangles, as well. The Cryo-electron tomography of tau fibril structures is very interesting, since they comprise not only AD-type tau misfolding but also CTE-like tau misfolding, suggesting a heterogenous origin of tau fibrils in AD brain. Since CTE tau pathology is frequently found in glial tangles as well as in neurofibrillary tangles, I was curious if there is any difference in origin of AD and CTE-like tau misfolding, such as one coming from neurons and another from glia. The difference in misfolding may also differentiate the pathological spread pattern of tau, which was shown previously in the tau seeding study in animal models (Narasimhan et al., 2017). 

    References:

    Narasimhan S, Guo JL, Changolkar L, Stieber A, McBride JD, Silva LV, He Z, Zhang B, Gathagan RJ, Trojanowski JQ, Lee VM. Pathological Tau Strains from Human Brains Recapitulate the Diversity of Tauopathies in Nontransgenic Mouse Brain. J Neurosci. 2017 Nov 22;37(47):11406-11423. Epub 2017 Oct 20 PubMed.

    View all comments by Tsuneya Ikezu

  5. I think this is an extremely important and exciting study. It would be great to see additional brains analyzed to learn more about earlier disease stages, in contrast to late-stage disease reported in this study, and to explore if these findings can be replicated in other AD patient brains. The presence of CTE-like tau filaments in AD brain, especially, warrants further study to address if these types of filaments are common in AD. In addition, it would be highly interesting to study other tauopathies as well, especially those where tau aggregation occurs in glial cells, since subcellular environments seem to impact tau filament structures. Another striking observation is the presence of extracellular vesicles (EVs), specifically in amyloid plaques. It is unclear whether EVs may just nonspecifically stick to pre-existing extracellular amyloid plaques, or if they play a role in seeding them. Previous research provided evidence that amyloid precursor protein (APP) and APP C-terminal fragments can be cleaved on the surface of EVs, thus releasing Aβ. In addition, the lipid composition of EV membranes has been implicated in facilitating aggregation of Aβ, supporting a possible role of EVs in seeding of plaques.

    My lab is also very interested in EVs and their role in neurodegenerative diseases. We recently described the first molecular pathology-specific biomarker for ALS and FTD, which is easily accessible from blood and will have important implications for future diagnosis and therapy in ALS and FTD spectrum disorders (Chatterjee et al., 2023). One major obstacle for therapy trials in FTD is the lack of biomarkers for antemortem detection of underlying molecular pathology. Without such biomarkers, patient recruitment to TDP-43 or tau-directed therapy trials has so far been limited to rare genetic FTD cases. Biomarkers, especially less invasive ones, are therefore urgently needed for selecting appropriate patients for clinical trials and to determine target engagement.

    In our work, we established a set of blood-based biomarkers that allow the reliable distinction between patients characterized by either tau or TDP-43 pathology. We show that plasma extracellular vesicles (EV) contain quantifiable amounts of TDP-43 as well as unfragmented tau which enables the quantification of 3-repeat (3R) and 4-repeat (4R) tau isoforms. We determined plasma EV TDP-43 levels and EV 3R/4R tau ratios in a large and deeply phenotyped cohort, comprising altogether 704 cases, including ALS as a TDP-43 proteinopathy, progressive supranuclear palsy (PSP) as a 4R tau predominant tauopathy, healthy controls and behavioral variant FTD (bvFTD) as a group which is associated with either tau or TDP-43 pathology.

    Importantly, we found that the combination of plasma EV TDP-43 and plasma EV 3R/4R tau ratios discriminates FTD cases with underlying TDP-43 pathology from those with tau pathology, with high sensitivity and specificity. This was confirmed by 63 cases with neuropathologically and/or genetically proven molecular pathology. Furthermore, high plasma EV TDP-43 levels distinguished cases with ALS, and very low tau ratios cases distinguished PSP, from other diagnostic groups, both with high diagnostic accuracies (AUC > 0.9). In addition, both markers strongly correlated with disease severity as assessed by multiple clinical and neuropsychological scales. Thus, plasma EV TDP-43 and tau ratio could not only aid molecular diagnosis in ALS, FTD, and PSP, but may additionally bear the potential of a marker to monitor disease progression and target engagement.

    References:

    Chatterjee M, Özdemir S, Fritz C et al. Plasma extracellular vesicle Tau isoform ratios and TDP-43 inform about molecular pathology in Frontotemporal Dementia and Amyotrophic Lateral Sclerosis. https://doi.org/10.21203/rs.3.rs-3158170/v1 (version 1) Research Square

    View all comments by Anja Schneider

  6. Our published paper now also includes much-improved sub-tomogram averaging of in-tissue tau filaments, using Warp/M and Relion, developed by Dimitry Tegunov at Genentech and Sjors Scheres at the Medical Research Council Laboratory for Molecular Biology, respectively. This yielded an 8.7 Å resolution, in-tissue structure of paired-helical fragments (PHF) from a single cluster composed of 136 tau filaments from a single tomogram. At this resolution we could trace the polypeptide backbone and unambiguously identify the protein fold.

    Mapping back this higher-resolution structure into the raw tomographic volume enabled us to identify the polarity orientation of each tau filament in tissue. Polarity orientation is an intrinsic property of all filamentous proteins. Because we found tau filaments are arranged in parallel clusters, the polarity of each filament could be oriented in one of two ways, top-to-bottom or bottom-to-top in the tomographic volume. We found that in the PHF only cluster, 114 filaments ran in one direction, but only 22 in the other. This highly skewed distribution of orientations was non-random. We think this indicates that the growth of one PHF filament has some degree of influence on the growth of the next, or that other factors within the cell control the polarity orientation of tau filaments.

    We also applied the improved sub-tomogram averaging pipeline to eight additional tau filament clusters from distinct in-tissue locations. This improved most of the maps, including a cluster of straight filaments that previously and misleadingly appeared CTE-like. We think that this highlights the importance of Warp/M-Relion and validation software for comparing sub-tomogram average maps with atomic models. To validate the fit of atomic models in our maps we collaborated with Randy Read, Cambridge Institute for Medical Research at the University of Cambridge, who developed a “log likelihood gain” scoring tool called em_placement that provides an absolute score of the fit of an atomic model into EM maps (Millán et al., 2023).

    References:

    Millán C, McCoy AJ, Terwilliger TC, Read RJ. Likelihood-based docking of models into cryo-EM maps. Acta Crystallogr D Struct Biol. 2023 Apr 1;79(Pt 4):281-289. Epub 2023 Mar 15 PubMed.

    View all comments by Rene Frank

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. Organization of Aβ Plaques and Tau Tangles Illuminated in AD Brain