Introduction
Microtubule-associated protein tau (tau) is intrinsically disordered and highly soluble when it is not bound to microtubules. The presence of hyperphosphorylated tau inclusions in the human brain pathologically defines a group of neurodegenerative diseases collectively referred to as tauopathies. The most prevalent tauopathy, AD, affects about 10% of the population over the age of 65 [
34]. In AD patients’ brains, cognitive decline and neuron death are closely associated with the increasing burden of aggregated tau filaments known as paired helical filaments (PHFs). Spatiotemporally, PHFs are found to spread throughout anatomically connected brain regions as the disease progresses. The spreading follows a stereotypical pattern beginning in the limbic system and is subsequently transmitted to the frontal temporal lobe and neocortex. Although the spatiotemporal pattern exhibits some variability, the spread of tau pathology through anatomically connected brain regions has been documented in other tauopathies as well [
8,
22,
23,
40]. Experimental evidence based on cell and animal models suggests that human brain-derived tau aggregates (tau seeds) are pathogenic, inducing the aggregation and fibrillization of endogenous, normal tau [
14,
15,
31]. Once induced into the pathological state, the newly formed tau aggregates recapitulate pathological features of human tau pathology including hyperphosphorylation, insolubility and cell-type specificity [
14,
21,
42,
43]. These models not only advance our insight into sporadic tau pathogenesis, but could also act as reliable biosensors for differentiating the various tau strains.
Different tau aggregates (referred to here as strains) may represent tau variants that lead to distinct tauopathies with different clinical and pathological phenotypes [
31]. Earlier evidence suggests that strain specificity appears to be conformation-dependent for different species of pathological tau [
14]. Biochemically, tau aggregates from different tauopathies show distinct patterns of resistance to protease digestion and/or denaturation, resulting in characteristic banding patterns of tau fragments on immunoblots [
31,
37]. Mass spectrometry data revealed the amino acid sequence of these digested fragments. The varied amino acid composition of the remaining fragments suggests that, in different tau strains, different regions of the pathological tau aggregates are exposed to proteolytic digestion [
37]. More recently, cryo-electron microscopy (Cryo-EM) studies confirmed that tau aggregates exhibit strain-specific structural heterogeneity at the ultrastructural level by revealing the core structures of AD, CBD, Pick’s disease (PiD) pathological tau filaments [
3,
4,
7,
44]. Interestingly, the structures of tau proto-filaments exhibit high uniformity within a given tauopathy, which is consistent with the strain-dependent activity observed in both sporadic cell culture and animal models of pathological tau transmission [
31].
Tau preformed fibrils (pffs), assembled in vitro in the presence of polyanionic inducers, share certain pathogenic features with human-derived tau filaments and potently induce tau aggregates in some tauopathy models. Past experiments have only demonstrated tau pff-induced pathogenesis in models that overexpress aggregation-prone, mutant human tau [
13,
17,
18,
21] and have failed to recapitulate these results in animal models that express full-length wild-type (WT) tau, which indicates that tau pffs lack some disease-relevant pathogenic features [
12,
14,
15]. The structural differences between tau pffs and human-derived tau aggregates provide a plausible explanation for the differences in activities, yet there is no direct evidence that connects the pathogenicity of different tau strains to their distinct structural conformations.
Though previous in vitro amplification experiments showed that pathological tau can be amplified from minute amounts and in a strain-dependent manner [
24,
28,
35], there is an unexplored connection between the amplified tau strains and their pathogenicities. In this study, we hypothesize that different tau strains, i.e., pathological tau species that adopt different conformations, can template conformation-dependent tau fibrillization by recruiting tau monomers and that, as a result, the strain-specific pathogenic features are conferred to and conserved in the fibrillized product. We standardized the single-cycle in vitro amplification reactions using human-derived tau strains from AD, CBD, and PSP patients’ brains as seeds and T40 as the substrate. The seeding process yielded AD-like (ADT40P1), CBD-like (CBDT40P1), and PSP-like (PSPT40P1) pathological tau filaments, which were biochemically analyzed to confirm their strain specificity. The pathogenicity of ADT40P1, CBDT40P1, and PSPT40P1 was quantitatively assessed in vitro using WT mouse hippocampal neuron cultures as well as in vivo using mouse models that express WT tau [
16].
Materials and methods
Animals
CD1 and C57BL6 wild type (WT) mice were purchased from Charles River. 5xFAD mice were purchased from Jackson Lab and bred with the B6/SJL F1 line for heterozygous offspring. 6hTau mice were generated by cross-breeding PAC-tau mouse line with PAC-E10 + 14 mouse line as described [
27,
45]. All animal care and experimental protocols were approved by the University of Pennsylvania’s Institutional Animal Care and Use Committee.
Purification of tau seeds
AD, CBD and PSP-tau were sequentially extracted from different regions of patient brains as previously described [
14]. All cases were screened to exclude the comorbidities of Lewy body and TDP-43 pathologies based on immunohistochemical staining (data not shown). The ELISA and immunoblot measurements of α-synuclein, amyloid beta and TDP43 protein concentrations are shown in Table
2. The whole middle frontal cortex except motor cortex was used for cortical extractions, while the posterior and middle temporal lobe was used for temporal lobe extractions. For AD cases, grey matter and white matter were separated using a surgical blade and only grey matter was used in the later extractions. Brain homogenate was prepared in nine volumes (v/w, ml/g) of PHF buffer (10 mM Tris, 10% sucrose, 0.8 M NaCl, 1 mM EDTA, pH 7.4) with 0.1% sarkosyl, proteinase inhibitor and phosphatase inhibitor in a glass dounce homogenizer and spun at 10,000 g for 10 min at 4 °C. Supernatant (sup 1) was collected, sarkosyl was added to 1% final concentration, incubated in a beaker with stirring for 1.5 h at RT, and followed by 150,000 g spin for 75 min at 4 °C. The pellet was collected, briefly washed with PBS to clean the myelin, and resuspended in PBS (pel 1). To remove the sarkosyl, pel 1 was spun at 150,000 g for 75 min and resulting pellet was collected as pel 2. For AD cases, pel 2 was resuspended in PBS, thoroughly sonicated and spun at 10,000 g for 10 min at 4 °C. The supernatant was collected as the enriched tau seeds. For PSP and CBD cases, pel 2 was used as tau seeds due to the relatively lower tau-burden in the tissue.
Expression and purification of recombinant tau
Recombinant tau proteins were expressed in BL21 (DE3) RIL
E. coli cells and purified by fast protein liquid chromatography following previous descriptions [
14]. The concentration of recombinant tau was determined using a nanodrop 1000 (ND-1000, Spectrophotometer). Heparin-induced tau pffs were made according to a previous protocol [
14].
In vitro seeding reactions
To deactivate proteolytic activity, tau seeds were first heated to 56 °C for 30 min in a thermocycler (PTC-200, Peltier Thermal Cycler) in the presence of a proteinase inhibitor cocktail (1% w/v pepstatin, leupeptin, TPCK, TLCK, and trypsin inhibitor, 0.1 M EDTA, in H2O) (1:100 v/v) and PMSF (50 mM). After the heat treatment, tau seeds were sonicated in a bath sonicator (Diagenode) for 20 min, with 30 s sonication and 30 s pause at the highest intensity. Tau seeds were mixed together with T40 to obtain a total tau concentration of 40 µM. For ADT40P1 and CBDT40P1 reactions, the ratio of tau seeds to monomer is 1:9. For ADT40P1*, the seeds/recombinant tau concentration ratio is 1:99. Next, 2 mM DTT was added into the solution. For PSPT40P1, the reaction was set up with 40 µM recombinant T40 with 0.2–2% of PSP-tau seeds depending on the tau concentration in the lysates. All the reactions were conducted in sterile phosphate-buffered saline (pH 7.4, without Mg2+ and Ca2+) in PCR tubes (PCR-02-C, Axygen Scientific) with 50–100 µl volume on a thermomixer (thermomixer C, Eppendorf) for 7 days at 37 °C if not specified. Day 0 samples were collected before agitation and kept frozen in a -80° C freezer until use.
Sedimentation assay
The sedimentation assay was conducted by mixing 1 µl of the sample with 19 µl of PBS (0.1% sarkosyl, w/v) and spun for 30 min at 100,000 g with an ultracentrifuge (Optima, Beckman Coulter) as previously described [
14]. The supernatant was removed and the pellet was suspended in 20 µl of PBS. Lysates were mixed with loading buffer for immunoblots.
Transmission and immuno-EM
EM was conducted as described previously [
14]. Negative staining of filaments for EM was performed with 2% uranyl acetate except as noted below. Immuno-EM was carried out using the AT8 antibody and the anti-myc antibody followed by anti-mouse and anti-rabbit secondary antibodies conjugated to 12 or 6 nm colloidal gold, respectively. Negative staining was performed in these experiments with 1% uranyl acetate. EM images were taken with a JEOL 1010 EM instrument. T40 monomers were treated with the same protocol to control for nonspecific staining. Quantification of EM was performed by ImageJ software.
Primary neuron cultures and transduction of tau fibrils
CD1 mouse cortices and hippocampi were dissected at embryo day 16–18 and dissociated with papain (Worthington Biochemical Corporation). Neurons were resuspended in neural basal medium (Gibco, 21,103) with 2% B27 (Gibco), 1 × Glutamax (Gibco) and 1 × penicillin/streptomycin (Gibco). Plates or coverslips were coated with poly-d-lysine (0.1 mg/ml, Sigma-Aldrich) in borate buffer (0.05 M boric acid, pH 8.5) overnight at room temperature. Cells were plated at a density of 50,000 cells/cm
2 for all types of plates. When plating, 5% FBS was added to the cell suspension. The plating medium was replaced with neural basal medium without FBS at day 1 in vitro (DIV1). At DIV7, the plating medium was replaced with conditioned medium containing 1:1 ratio of old and fresh medium. Tau fibrils were diluted in PBS to the desired concentrations, added on top of the cells and incubated for 14 days until DIV 21. The dose of tau fibrils was scaled based on cell densities (μg tau/10
6 cells) per different plate. For activity testing, 1 μg tau/10
6 cells was used for 100% seeds (100% pathogenic tau), 0.1 μg tau/10
6 cells for 10% seeds control and 0.01 μg tau/10
6 cells for 1% seeds control unless otherwise specified. All the amplified tau variant doses were based on the total tau concentration at the start of the in vitro reaction. For PSP seed controls, the seeds were diluted to match the reaction conditions and transduced onto the neurons. To compare the ultracentrifuge fractionated pellet of the amplified tau variants to the seeds, the tau concentrations were first estimated using immunoblots (Fig.
3c) and then transduced on neurons with the same dose.
Immunocytochemistry
At DIV 21, cells in 24 well plates with coverslips were washed with PBS once and fixed with ice-cold methanol for 20 min in a − 20 °C freezer. For 96 and 384 well plates, cells were washed four times with PBS in a plate washer (BIOTEK), extracted with 1% hexadecyltrimethylammonium bromide (HDTA, Sigma-Aldrich) for 10 min and fixed for 10 min with 4% paraformaldehyde and sucrose in PBS. Fixed cells were stained with primary antibodies overnight at 4 °C and Alexa Fluor-conjugated secondary antibodies (Thermo Fisher Scientific) for 2 h at room temperature. Photomicrographs were obtained with a light microscope (Eclipse Ni, Nikon) and quantified using HALO software as described previously [
14]. Confocal photomicrographs were taken with a Leica DMI 6000 microscope coupled with a white-light laser.
Stereotaxic injection of pathogenic tau fibrils
Mice were anesthetized with ketamine/xylazine/acepromazine and fixed on a stereotaxic frame (David Kopf Instruments). 5xFAD and WT mice were unilaterally double injected with tau fibrils at bregma: − 2.5 mm, lateral: + 2 mm, depth: − 2.4 mm and − 1.4 mm from the skull at right dorsal hippocampus and cortex. Two micrograms of tau was injected for the 100% seeds controls (2 μg AD-tau) and ADT40P1 (0.2 μg AD-tau + 1.8 μg T40 with agitation for 7 days) or 10% seeds control (0.2 μg of AD-tau + 1.8 μg T40 without agitation). 6hTau mice were bilaterally injected (bregma: − 2.5 mm, lateral: ± 2 mm, depth: − 2.4 mm from the skull). The right hemisphere was used for histology. The left hemisphere was used for biochemistry.
Cells were washed once with PBS and harvested in PBS containing 1% sarkosyl and proteinase inhibitor cocktail. The lysates were incubated on ice for 15 min and then spun at 100,000 g for 30 min at 4 °C. The supernatant was kept as the soluble fraction, and the pellet was suspended in PBS as the insoluble fraction. For mouse brain tissues, hippocampi were dissected after perfusion with PBS at a flow rate of 1.5 ml/min for 20 min and homogenized in nine volumes of PHF extraction buffer containing 1% Triton-X100. Lysates were spun at 100,000 g for 30 min at 4 °C. The supernatant was collected as Triton-soluble fraction (TX-S), and the pellet was suspended in distilled water containing 1% sarkosyl, incubated at room temperature for 1 h and then spun at 100,000 g for 30 min at 4 °C. The supernatant was collected as the sarkosyl-soluble fraction (Sark-S) and the pellet saved as the sarkosyl-insoluble fraction (Sark-P). Fifteen micrograms of total proteins from TX-S and Sark-S was loaded on the immunoblots. Sark-P was loaded at a 10:1 ratio to the respective TX-S fractions. The immunoblots were generated as previously described [
13]. Primary antibodies were incubated overnight at 4 °C followed by secondary incubation of 2 h at room temperature. Photomicrographs were taken using an imager (Li-Cor Biosciences). The optical densities were measured using ImageJ Software. For dot blot studies, AD-tau, ADT40P1 and T40 samples were diluted to the desired concentrations in Tris-buffered saline (TBS) and loaded on 0.4 mm PVDF membranes and blocked with 5% milk. Primary and secondary antibodies (Thermo Fisher Scientific) were incubated following the same protocol as for immunoblots.
Heatmaps of tau pathology
Heatmaps of the distribution of AT-8 tau pathology were generated as previously described [
15]. Each region of the brain was scored for the presence of tau pathology between 0 and 3, given 0 is completely free of tau pathology and 3 is the most abundant tau pathology. The semiquantitative scores of AT-8 tau pathology were obtained in a blinded manner.
Histology and immunofluorescence (IF) staining
Mouse brains were dissected after perfusion with PBS at a flow rate of 1.5 ml/min for 20 min. The brains were immersion fixed with 10% neutral buffered formalin (NBF) and processed following a previously described protocol [
15]. Six-micrometer-thick brain sections were stained with different primary antibodies and developed using a polymer horseradish peroxidase detection kit (BioGenex). For immunofluorescence, the brain sections were incubated with 3R (RD3) and 4R (4Rtau) tau-specific antibodies overnight at 4 °C followed by a 2-h incubation of Alexa Fluor-conjugated secondary antibody (Thermo Fisher Scientific). Autofluorescence was quenched using a 0.3% Sudan black solution.
Proteinase K digestion
ADT40P1, CBDT40P1 and PSPT40P1 pellets were fractionated using an ultracentrifuge at 100,000 g for 30 min and suspended in PBS. Tau concentrations were estimated using immunoblots. 0.5 μg of tau variants was mixed with 0.0005% (w/v) proteinase K in PBS buffer and incubated at 37 °C for 30 min. The reactions were stopped by boiling the samples at 100 °C with 1 × loading buffer for 10 min. The same concentrations of T40 and hepT40 were used as controls.
WND-CHARM analysis of tau pathology
Supervised classification was performed for testing of the morphological differences of tau pathologies induced by different tau strains. Top 15% of the images features were used for training the classifier. For each time of the test, a random set of 70% of the images were used to train the classifier to predict AD vs PSP vs CBD using the image features. The other 30% of the images were used for the classification. The test was performed 100 times independently. Twenty-five percent of downsample was used in the analysis. Command line is: wndchrm test -f0.15 -r#0.70 -n100 -d25 -N3 –mls. Images from four biological repeats were used for the analysis.
Antibody dilution
All antibodies were diluted according to the supplementary table, online resource.
Statistical analyses
Unpaired t tests were performed to compare two experimental groups. For comparisons of more than two groups, one-way ANOVA and two-way ANOVA was performed, followed by Tukey post hoc test. All the statistics were performed using Prism software (GraphPad Software, Inc). Statistical significance was reached when p ≤ 0.05. If not specified, data are presented as the mean ± standard deviation.
Discussion
In the present study, we demonstrated that patient-derived tau (AD-tau, CBD-tau and PSP-tau lysates) can seed monomeric recombinant T40 in vitro, resulting in strain-dependent amplification of the pathogenic tau conformers. The amplified tau conformers potently induce strain-dependent tau pathology in both cell and animal models without requiring the overexpression of WT or mutant tau. In vitro assays and models demonstrate that the amplified material recapitulates the pathogenic features of its seeds, including insolubility, filamentous ultrastructure, and intraneuronal distribution. In vivo models demonstrate that intracerebral injection of amplified materials induces tau pathology with strain-specific, morphological and spatial–temporal features, including specific isoform recruitment and total tau-burden. This study provides the first evidence that the strain-specific pathogenic features of human-derived tau aggregates can be conferred to monomeric tau during an in vitro amplification reaction.
Previous in vitro tau seeding reactions used tau fragments that only contained the microtubule-binding domains [
9] or used aggregation-prone mutant tau as the substrate to promote tau fibrillization in the reaction [
5,
39]. These studies provided an important foundation for the in vitro fibrillization of tau using brain-derived tau seeds extracted from pathologically diagnosed tauopathy cases. The real-time quaking-induced conversion (RT-QUIC) and protein misfolding cyclic amplification (PMCA) assays demonstrated an increase in beta-sheet fibrillization, as measured by the huge increase (10
7- to 10
10-fold) in fluorescent signal of a beta-sheet binding compound [
24,
29]. However, past attempts at in vitro fibrillization yielded synthetic tau strains with highly variable bioactivities and fidelity and several caveats make these fibrillization reactions unsuitable for authentic, disease-relevant amplification of tau seeds. First, these reactions used polyanionic inducers, which lead to the formation of nonspecific, preformed tau fibrils and contributed to the bioactive variability of the assembled fibrils. Second, the use of mutant tau, which has been shown to be especially prone to aggregation, also interferes with the properties of the fibrils [
36]. Third, the thioflavin-based fluorescent signal has been shown to correlate poorly with the biological activities of tau conformers [
1]. The bioactivity of the amplified material from previous studies remains largely unknown, and so far there is no convincing evidence supporting that the pathogenicity of tau strains could be amplified and maintained in multiple cycles.
In this study, we chose to use T40, a full-length 2N4R tau isoform, as the monomeric substrate because it contains all the domains of mature tau in the adult human brain. Importantly, T40 is the common tau isoform present in AD-tau, CBD-tau and PSP-tau. The single-cycle fibrillization reaction did not include any known polyanionic inducers or cofactors to eliminate spontaneous de novo tau fibrillization. Additionally, our use of physiological in vitro and in vivo models with endogenous mouse tau expression provides the necessary resolution to differentiate between the pathogenic conformers. These experimental systems allow us to generate bioactive tau strains in vitro and accurately assess their strain-specific pathogenicities.
Of note, we did not observe evidence that the amplified fibrils are active in secondary in vitro amplification reactions. Two hypotheses could explain the inertia of the amplified products in subsequent in vitro amplification reactions. One hypothesis is that the initiation of tau aggregation in human and mouse neurons is driven by pathogenic cofactors, which are needed to induce the soluble tau monomers into an intermediate state that allows them to interact with the seeds. While these cofactors could be present in the disease brain lysates during the first cycle of amplification, as well as in the mouse brains and in the cultured neurons, they may get diluted in secondary cycles of in vitro amplification. Previous studies have also hypothesized that cofactors could affect tau behavior in in vitro reactions [
6,
38]. This hypothesis is also consistent with the observation that in the PMCA and QUIC assays, polyanionic factors must be present for multiple cycles of amplifications. To address this question, human tau seeds with higher purity will be needed to further assess the role of these cofactors in tau pathogenesis.
The second hypothesis is that posttranslational modifications of human tau fibrils enable the human-derived tau seeds to interact with recombinant tau and induce fibrillization. The T40 monomers, and thus, the resulting amplified material, lack the active sites necessary for successive seeding activity due to the absence of posttranslational modification. In human disease brains, tau fibrils are known to be extensively posttranslationally modified and, according to this hypothesis, these posttranslational modifications may be necessary for the fibrils to interact with tau monomers. In the first round of in vitro amplification, the T40 monomers can interact with human-derived tau seeds to form fibrils. However, during this initial amplification reaction, the active tau seeds become saturated with T40 fibrils, which lack posttranslational modifications. Thus, the resulting amplified fibrils are inert and can no longer induce fibrillization in vitro. Despite their inertia in subsequent in vitro amplification reactions, the amplified materials are potent in neurons and mouse brains because the amplified fibrils are reactivated by posttranslational modifications following endocytosis. This hypothesis is consistent with our observation that, at the endpoint of the amplification reaction, some of the T40 monomers remain unfibrillized (Supplementary Fig. 1a, online resource). Experimental evidence of posttranslational modifications and their effect on the fibrillization of tau will be essential to further address this hypothesis.
The quality of the extracted tau seeds is crucial for the in vitro amplification reactions. It is well documented that tau can bind to anions or lipids in the cells, which could lead to variability in quantifying the amount and/or activities of the amplified material. To enrich for brain-derived tau seeds in our study, we chose tauopathy cases at late stages of the disease where tau-burden is high in many brain regions. The presence of copathology in a primary tauopathy case may lead to the generation of heterogeneous tau seeds, so the extracted brain regions were also histochemically screened for other proteinopathic comorbidities, such as a-synuclein and TDP-43 inclusions. Brain tissue containing comorbid pathology was not selected for extraction, which enabled a higher yield and greater enrichment of high-quality tau seeds. Furthermore, using this approach to human tissue selection, we were able to experimentally determine that tau seeds isolated from different individuals with the same type of tauopathy are comparable in terms of their biochemical properties as well as pathogenicity in cell and mouse models. Despite our thorough screening process, we cannot exclude the possible existence of subspecies of tau seeds that could be underrepresented by the enrichment procedure. A systematic investigation of different brain regions, disease courses, and cell types will be needed to address the possible heterogeneity of human-derived tau seeds.
Additionally, the CBD and PSP lysates were much lower in their concentration and purities compared to AD-tau, which is due to a lower tau-burden in CBD and PSP patient brains. Despite the relatively low yield and quality of the CBD and PSP lysates, we did observe a comparable increase in the activity of the amplified material, which was consistently a two- to threefold increase in activity compared to the seeds control across the different tau strains. The consistency of amplification across different lysates strongly indicates that tau seeds, but not other insoluble molecules in the lysates, are recruiting the recombinant tau monomers. However, we have not yet to rule out the possibility that insoluble substances in the lysates will interact with tau and cause nonspecific aggregation of tau monomers. Nonetheless, to effectively increase the amplification efficiency further, a higher purity of tau seeds may be necessary.
Additionally, our results indicate that multiple isoforms of tau are not required for the faithful fibrillization of tau in vitro, but we have yet to explore whether the presence of multiple tau isoforms could promote fibrillization in the seeding reaction. In a previous study, we tested seeding reactions using recombinant human tau with a 1:1 ratio of 3R and 4R tau isoforms and found no significant increase compared to the single isoform reactions [
14], yet we cannot rule out the possibility that a different combination of monomers could lead to different outcomes of the seeding reaction. Further experiments to address these possible variables could also optimize the efficiency of the fibrillization reaction.
In human brains, the transmission of tau pathology occurs in mature neurons with intact neuroanatomical connectomes. However, since pathological tau is also found in oligodendrocytes and astrocytes, these glial cells could play an important role in the development of tau pathology and its transmission in the disease state. Thus, it is essential to use intact animals for in vivo studies of tau pathology transmission. In this study, we confirmed the pathogenicity of ADT40P1 in vivo using both 5xFAD transgenic mice and WT mice. The 5xFAD mice overexpress human mutant APP and PSEN1, which leads to the accumulation of amyloid-beta plaques in the adult mouse brain [
32]. However, both the 5xFAD and WT models express endogenous mouse tau at physiological concentrations and do not develop tau pathology without injection of human tau seeds. Synthetic tau pffs are not capable of inducing a significant amount of tau pathology in either the 5xFAD mouse or WT mouse brains [
14]. Intracerebral injection of human tau seeds into 5xFAD mice (at least 4 months of age) will lead to the formation of neuritic plaque tau pathology at the one-month post-injection interval [
15]. Thus, the 5xFAD mice are a robust and reliable model for quantitative measurement of tau pathology with the resolution required to differentiate human tau seeds from artificially assembled tau pffs.
In this study, the increased tau pathology observed in 5xFAD mice following intracerebral injection of ADT40P1 versus a 10% AD-tau seeds control is consistent with the in vitro data from the neuron culture assay, confirming that ADT40P1 is pathogenic and potently induces tau pathology in vivo. ADT40P1 also induced a significant amount of tau pathology in WT mice, further validating its disease-relevant pathogenicity. ADT40P1 induced neuritic plaque tau pathology indistinguishable from that induced by extracted AD-tau, suggesting that ADT40P1 could be used in lieu of AD-tau, which is a very limited resource, for further in vivo studies of tau pathogenesis.
Prior studies have demonstrated that different tau strains are composed of different tau isoforms [
37]; however, there has been no direct evidence demonstrating that the tau isoform composition in the fibrils dictates strain-specific bioactivity. We hypothesize that the structural conformation of the fibrils, rather than the isoform composition, determines the bioactivity of different tau strains. To demonstrate this hypothesis, we interrogated the activity of ADT40P1 and CBDT40P1 in 6hTau transgenic mice as an additional in vivo model. The 6hTau mice express all six isoforms of human WT tau on a null mouse tau background, mimicking the isoform expression patterns in the human adult brain and providing a human-like tauopathy model to elucidate the isoform recruitment of different tau strains [
27]. With WT human tau overexpression, the 6hTau mice are free of tau pathology throughout their lifespan. However, intracerebral injection of human tau seeds in 6hTau mice induced strain-specific tau pathology that correlated with the tau isoform composition of the seeds [
45]. We normalized the total tau pathology by scaling the dose of the injected material so that we could perform a direct comparison of different tau isoform components in the induced tau aggregates. We observed that the tau aggregates in ADT40P1- and AD-tau-injected cohorts had the same ratios of 3R to 4R tau isoforms despite the fact that the ADT40P1 fibrils are primarily composed of 4R tau. CBDT40P1 and CBD-tau injected cohorts had primarily 4R tau aggregates. Together with the PK digestion data, these findings indicate that the pathogenicity of amplified material was inherited through the transmission of the abnormal tau conformation, rather than the isoform composition. The data also suggest that the isoform recruitment is a feature of strain-specific bioactivity, rather than a causal agent for the formation of different tau seeds. Here, we demonstrated that a single tau isoform converted by different tau strains into a pathological conformation is sufficient to recapitulate the signature bioactivity, including strain-specific isoform recruitment, of pathological tau in vivo.
Our study demonstrates that the pathogenicity and biophysical properties of human-derived tau seeds are conferred to and conserved in the products of in vitro amplification reactions. These findings not only demonstrate an experimental framework for understanding disease-specific tau fibrillization, but are also a proof-of-concept that, under proper conditions, in vitro amplification reactions yield pathogenic material that recapitulates the bioactivity of human-derived tau seeds that can dramatically increase the availability of pathogenic tau material for future studies of tau pathogenesis.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.