Extracellular monomeric tau protein is sufficient to initiate the spread of tau protein pathology

Abstract

Understanding the formation and propagation of aggregates of the Alzheimer disease-associated Tau protein in vivo is vital for the development of therapeutics for this devastating disorder. Using our recently developed live-cell aggregation sensor in neuron-like cells, we demonstrate that different variants of exogenous monomeric Tau, namely full-length Tau (hTau40) and the Tau-derived construct K18 comprising the repeat domain, initially accumulate in endosomal compartments, where they form fibrillar seeds that subsequently induce the aggregation of endogenous Tau. Using superresolution imaging, we confirm that fibrils consisting of endogenous and exogenous Tau are released from cells and demonstrate their potential to spread Tau pathology. Our data indicate a greater pathological risk and potential toxicity than hitherto suspected for extracellular soluble Tau.

Keywords: Alzheimer Disease; Amyloid; Endocytosis; Fluorescence Lifetime Imaging Microscopy; Propagation; Protein Aggregation; Superresolution Microscopy; Tau.

FIGURE 1.

The fluorescence lifetime of Alexa Fluor 488-labeled K18* and hTau40* reports on the structural conformation of the protein in vitro and in vivo. A, K18* (i) and K18*-488 (ii), each incubated at 10 μm with 2.5 μm heparin for 6 h prior to transmission electron microscopy analysis, form similar fibrils. Similarly, hTau40* (iii) and hTau40*-488 (iv), each incubated at 10 μm with 2.5 μm heparin for 240 h prior to transmission electron microscopy analysis, form similar fibrils. Scale bars = 100 nm. B, the fluorescence lifetime of 10% K18*-488 (1 μm in cell culture medium) and 1% hTau40*-488 (20 μm in BES buffer) was either measured immediately (i and iii for K18* and hTau40*, respectively) or after incubation with 4:1 Tau:heparin for 24 h at 37 °C (ii and iv for K18* and hTau40*, respectively). Prior to inducing aggregate formation by heparin, the excited-state lifetime of K18*-488 lies at 3664 ± 25 ps (i) and that of hTau40*-488 at 3657 ± 18 ps (iii). After heparin treatment, the fluorescence lifetime of K18*-488 drops to 3262 ± 71 ps (ii), whereas that of hTau40*-488 drops to 3552 ± 26 ps (iv). Scale bars = 10 μm. C, 1 μm 10% K18*-488 or hTau40*-488 were added to the extracellular medium of SH-SY5Y cells, and their fluorescence lifetimes were analyzed after 8 h incubation (i and iv for K18* and hTau40*, respectively). The fluorescence lifetimes of the extracellularly added K18*-488 and hTau40*-488 remained at 3658 ± 16 ps and 3686 ± 16 ps, respectively, indicating that Tau does not aggregate in the absence of an aggregation-inducing agent, such as heparin, in the extracellular space. Confocal (ii and v for K18* and hTau40*, respectively) and TCSPC images (iii and vi for K18* and hTau40*, respectively) of SH-SY5Y cells, which were incubated for 8 h with 1 μm 10% K18*-488 or hTau*-488 and then trypsin-washed prior to imaging, are shown. The fluorescence lifetimes of K18*-488 and hTau40*-488 dropped to 3282 ± 72 ps and 3154 ± 56 ps, respectively. Scale bars = 10 μm. D, bar diagram displaying the mean fluorescence lifetime values of the different samples measured. Error bars represent S.D. One-way analysis of variance was performed for K18* (F (3, 102) = 616.7, p < 0.0001) and hTau40* (F (3, 74) = 980.5, p < 0.0001). ****, p < 0.0001; n.s., not significant.

FIGURE 2.

Glycosaminoglycans are not involved in Tau uptake by cells. A, PgsA-745 cells, which are deficient in glycosaminoglycans, were incubated with 1 μm 10% K18*-488. At t = 0, the fluorescence lifetime in the extracellular space was 3826 ± 36 ps, corresponding to soluble Tau. After 8 h, the cells were tryspin-washed, and the fluorescence lifetime of internalized K18*-488 was 3448 ± 54 ps, indicative of aggregation. Scale bars = 10 μm. B, bar diagram displaying the mean fluorescence lifetime values of the different samples measured. Error bars represent S.D. Unpaired Student's t test statistical analysis was performed. ****, p < 0.0001.

FIGURE 3.

K18*-488 is taken up by endocytosis. A, SH-SY5Y cells were incubated with 1 μm 10% K18*-488 (green) for 24 h, and the lipid marker FM 4-64 (red) was added during the last 15 min of incubation. The cells were imaged by confocal microscopy. Colocalization between FM 4-64 and K18*-488 is displayed in yellow. Scale bar = 5 μm. B, SH-SY5Y cells were incubated with 1 μm 10% K18*-488 (green) either at 37 °C (left panel) or 4 °C (right panel). After 1 h, both dishes were trypsin-washed and imaged by confocal microscopy. The images display more internalized K18*-488 when the cells are incubated at 37 °C rather than at 4 °C. Scale bars = 5 μm. C, K18*-488 was incubated for 24 h at 100 μm in cell culture medium and compared with K18*-488 incubated at 1 μm in cell culture medium. The fluorescence lifetimes measured were 3765 ± 42 ps and 3664 ± 25 ps, respectively. Error bars represent S.D. D, K18*-488 was incubated at 1 μm (10% labeled) in cell culture medium at pH 4.7. hTau40* was incubated at 10 μm (1% labeled) in BES buffer at pH 4.7. The fluorescence lifetimes of soluble (3824 ± 22 ps) and aggregated K18*-488 (3268 ± 31 ps) and of soluble (3765 ± 53 ps) and aggregated hTau40*-488 (3409 ± 189 ps) indicate that low pH is sufficient to induce aggregation of both K18* and hTau40* in vitro in the absence of heparin. Error bars represent S.D. Unpaired Student's t test statistical analysis was performed. ****, p < 0.0001. E, Alexa Fluor 488 was dissolved in cell culture medium to 100 nm (pH 4.7 or 7.4), and the mean fluorescence lifetime was measured to be 3889 ps ± 28 ps and 3844 ± 14 ps, respectively. Error bars represent S.D.

FIGURE 4.

K18* seeds endogenous Tau aggregation. A, confocal and TCSPC images of SH-SY5Y cells incubated for 8 h with K18*-488 or hTau40*-488, respectively, trypsin-washed, and incubated in Tau-free medium for 64 h. Scale bars = 10 μm. B, corresponding bar diagram displaying the mean fluorescence lifetimes of K18*-488 (2624 ± 91 ps) and hTau40*-488 (2779 ± 132 ps), with dashed lines indicating fluorescence lifetimes of monomeric and endocytosed K18*. Error bars represent S.D. One-way analysis of variance was performed for K18* (F (2, 77) = 2080, p < 0.0001) and hTau40* (F (2, 56) = 352.5, p < 0.0001). ****, p < 0.0001. C, total Tau content was analyzed by Western blotting after sequential solubilization in Triton, sarkosyl, and SDS, respectively. Lane 1, cell lysates of cells incubated for 8 h with K18*. Lane 2, cell lysates of cells incubated for 8 h with K18*, trypsin-washed, and incubated for 64 h in Tau-free medium. Lane 3, cell lysates of age-matched control cells. Single arrowheads indicate monomeric K18*, double arrowheads point to K18* dimers, and arrows point to endogenous Tau. Tubulin was used as a loading control in the Triton-soluble fraction. D and E, 1 μm 10% K18*-488 was incubated for 24 h with 0.01% trypsin, and the fluorescence lifetime of K18*-488 was determined either prior to (i) or after (ii) trypsin incubation. The bar diagram represents the mean fluorescence lifetime of K18*-488 before (3664 ± 25 ps) or after trypsin incubation (3940 ± 50 ps). Scale bars = 10 μm. Error bars represent S.D. Unpaired Student's t test statistical analysis was performed. ****, p < 0.0001. F, SH-SY5Y cells were incubated with 1 μm 10% K18*-488, and the mean fluorescence lifetime of the fluorophore was recorded under the following conditions. Navy blue, extracellular space; medium blue, intracellular space; light blue: following uptake, trypsin-wash, and incubation in Tau-free medium. Error bars represent S.D.

FIGURE 5.

Exogenously added and internalized Tau is released by cells. A, SH-SY5Y cells were incubated with either K18*-488 or hTau40*-488 for 72 h, trypsin-washed, and incubated in Tau-free medium for 96 h prior to imaging. By analyzing the cell culture medium of these cells by TCSPC, the following fluorescence lifetimes for K18*-488 and hTau40*-488 were measured, respectively: 2807 ± 30 ps (i) and 2749 ± 49 ps (ii). Scale bars = 10 μm. B, the corresponding bar diagram represents the mean fluorescence lifetimes observed for released Tau. The dotted lines represent the mean fluorescence lifetime of monomeric and endocytosed K18*-488. Error bars represent S.D. One-way analysis of variance was performed for K18* (F (2, 62) = 1026, p < 0.0001) and hTau40* (F (2, 47) = 848.5, p < 0.0001). ****, p < 0.0001. C, the cell culture medium of cells treated with K18*-647 for 72 h, trypsin-washed, and incubated in Tau-free medium for 96 h was immunostained with the TAUY9 antibody prior to analysis by two-color dSTORM. The image displays K18*-647 fibrils (red) that are intertwined with endogenous Tau (green). Scale bar = 1 μm.

FIGURE 6.

Released K18* seeds induce aggregation of full-length hTau40. A, culture medium containing released unlabeled K18* was incubated with 1 μm 10% hTau40*-488 for 48 h at 37 °C, and the mean fluorescence lifetime of hTau40*-488 was monitored before (i) and after (ii) incubation. Scale bars = 10 μm. B, bar diagram representing the mean fluorescence lifetime of hTau40*-488 prior (3735 ± 31 ps) and after incubation (3388 ± 33 ps) with culture medium containing released unlabeled K18*. Error bars represent S.D. Unpaired Student's t test statistical analysis was performed. ****, p < 0.0001. C, hTau40*-647 was incubated with soluble unlabeled K18* (i, ii, and iii) or with culture medium containing released K18* for 48 h at 37 °C (iv, v, and vi) prior to dSTORM imaging. The fluorescence images of induced fibrils are displayed (i and iv). Color images show the corresponding (ii and v) and zoomed (iii and vi) superresolved images of the fibrils formed obtained by dSTORM. hTau40*-647 incubated with medium containing released K18* forms fibrillar structures, whereas hTau40*-647 incubated with soluble K18* does not display any fibril formation. Scale bars = 1 μm.

FIGURE 7.

Exogenously added and internalized Tau propagates from cell to cell. A, the culture medium containing either released K18*-488 or released hTau40*-488 was transferred onto fresh cells and incubated for an additional 72 h. Re-uptaken K18*-488 displays a fluorescence lifetime of 2637 ± 40 ps (ii), and i displays the corresponding confocal image. Re-uptaken hTau40*-488 displays a fluorescence lifetime of 2575 ± 51 ps (iv), and iii displays the corresponding confocal image. Scale bars = 10 μm. B, corresponding bar diagram representing the mean fluorescence lifetimes observed for re-uptaken Tau. The dotted lines represent the mean fluorescence lifetime of monomeric and endocytosed K18*-488. Error bars represent S.D. One-way analysis of variance statistical analysis was performed for K18* (F (2, 72) = 2995, p < 0.0001) and hTau40* (F (2, 46) = 1120, p < 0.0001). ****, p < 0.0001.

FIGURE 8.

Monomeric Tau initiates the misfolding cycle. Shown is a proposed model of the spread of Tau pathology. Intracellular monomeric Tau is released into the extracellular space through processes involving neuronal death or exocytosis, which is supported by elevated levels of Tau measured in the cerebrospinal fluid of AD patients. Monomeric Tau (1) can now directly be taken up by surrounding neurons through endocytosis (2). The intravesicular environment, such as present in endo- or lysosomes, then promotes the nucleation of endocytosed Tau (3). Our model does not currently address the question whether Tau leaves the intravesicular compartment or not. Thus, exogenous Tau can either be released into the cytoplasm (4), where it can seed aggregation of endogenous healthy Tau (5) or encounter endogenous Tau targeted for intravesicular degradation (4′) and seed the latter (5′). Heterogeneous aggregates are consequently released into the extracellular space either through not yet defined pathways such as exosomal release or simply by cell death (6) or by lysosomal exocytosis (6′). Thus, the propagation of misfolded Tau can now proceed as suggested previously (7).