Viral Delivery of Non-Mutated Human Truncated Tau to Neurons Recapitulates Key Features of Human Tauopathy in Wild-Type Mice

Abstract

Background:

Neuronal accumulation of hyperphosphorylated and truncated tau aggregates is one of the major defining factors and key drivers of neurodegeneration in Alzheimer’s disease and other tauopathies.

Objective:

We developed an AAV-induced model of tauopathy mediated by human truncated tau protein without familial frontotemporal dementia-related mutations to study tau propagation and the functional consequences of tau pathology.

Methods:

We performed targeted transductions of the hippocampus or entorhinal cortex in adult mice followed by histological analysis to study the progression of hippocampal tau pathology and tau spreading. We performed behavioral analysis of mice with AAV-induced hippocampal tau pathology.

Results:

AAV-induced hippocampal tau pathology was characterized by tau hyperphosphorylation (AT8 positivity), sarkosyl insolubility, and the presence of neurofibrillary tangles. AAV-induced tau pathology was associated with microgliosis and hypertrophic astrocytes in the absence of cognitive deficits. Additionally, the co-expression of mCherry fluorescent protein and human truncated tau enabled us to detect both local spreading of human tau and spreading from the entorhinal cortex to the synaptically connected dentate gyrus.

Conclusion:

Targeted delivery of AAV with truncated tau protein into subcortical and cortical structures of mammalian brains represents an efficient approach for creating temporally and spatially well-defined tau pathology suitable for in vivo studies of tau propagation and neuronal circuit deficits in Alzheimer’s disease.

Keywords

Adeno-associated virus Alzheimer’s disease dementia mice tau protein tauopathies

INTRODUCTION

Tauopathies are a heterogeneous class of important neurodegenerative disorders such as Alzheimer’s disease (AD), Pick’s disease, and progressive supranuclear palsy, all characterized by intracellular accumulation of aggregated tau protein [1]. Currently the most common tauopathy and the leading cause of dementia, AD affects 45 million people worldwide, and this number is expected to rise to 130 million by 2050 [2].

Tau pathology starts to develop when tau protein adopts a misfolded conformation that leads to aggregation and incorporation of physiological tau monomers—a process known as templated misfolding or “seeding”. Continued seeding leads to formation of highly structured insoluble fibrils which accumulate intracellularly as neurofibrillary tangles (NFTs) [3]. Tau is a highly soluble protein and it is currently unclear what causes the energetically unfavorable event of aggregation. However, post-translational modifications of tau such as phosphorylation and truncation can lead to abnormal conformations which are prone to aggregation [4]. Misfolded tau proteins can also transfer between neurons and seed physiological tau in other cells, thereby potentially spreading the pathology to healthy brain regions [5, 6]. This process is thought to underlie the stereotypical progression of tau pathology throughout the brain in AD patients [7]. Tau pathology is additionally accompanied by chronic neuroinflammation, with reactive microglia and astrocytes [8]. The entire pathological process ultimately culminates in dysfunction of neuronal networks, resulting in progressive cognitive impairment [9].

Animal models of human tauopathies using adeno-associated viruses (AAV) have emerged recently as flexible and inducible approaches offering several key advantages over traditional transgenic animal models [10]. Expression of tau can be induced in any brain region of interest, leading to fewer undesirable off-target effects (e.g., paralysis caused by expression in the brain stem), and allows for the study of tau spreading [11]. Incorporating cell type-specific promotors allows for inducing tau expression only in the target cell type. AAV-based models do not require specific genotyping and tau pathology can be induced on various genetic backgrounds. However, AAV-based models using full-length tau protein have also presented with several drawbacks, which are shared with traditional transgenic tauopathy models. Most notably, human tau shows only very limited propensity for aggregation. This has led to the use of tau mutations that are only found in familial frontotemporal dementia, but not in the majority of patients suffering from sporadic tauopathies such as late-onset AD [10].

Tau fragments cleaved at both the N- and C-terminals are abundantly found in the brains of patients with AD, Pick’s disease, and progressive supranuclear palsy, and truncation may therefore play a role in the pathogenesis of tauopathies [12 –15]. The N-terminal is mostly absent from seed-competent tau derived from AD brains but present in transgenic animals, indicating that tauopathy mouse models that overexpress tau with FTD mutations might not recapitulate the biochemical characteristics of human tauopathies [15 –17]. A toxic soluble tau fragment—previously associated with cognitive impairment in AD—is also associated with cognitive impairment in Huntington’s disease and Lewy body diseases such as Parkinson’s disease and dementia with Lewy bodies [18 –20]. Truncation likely exposes the repeat domains of tau that are normally covered by the C-terminal, thereby increasing the propensity for aggregation [21]. Inhibition of tau cleavage by the proteases asparagine endopeptidase or calpain in P301S mice led to decreased neurofibrillary pathology and rescue of cognitive deficits [22, 23]. Promoting pathways that lead to increased cleavage of tau was associated with accumulation of aggregated tau [14, 24]. Overexpression of truncated tau in both transgenic mice and rats led to neurofibrillary pathology and functional deficits [25 –28]. These studies therefore indicate that truncated tau protein may be critically involved in the pathological process of tau aggregation.

Here we present the characterization of an AAV-based animal model that induces expression of non-mutated human truncated tau, AAV9-Synapsin-Tau(151–391/4R)-mCherry (AAV-hSyn-htTau). AAV-hSyn-htTau injected into the brains of wild-type mice led to accumulation of insoluble hyperphosphorylated tau, NFTs, seeding of endogenous mouse tau, and spreading of human truncated tau to healthy neurons. Developing truncated tau-mediated pathology was associated with mild microgliosis and hypertrophy of astrocytes in CA1. Targeted application of AAV-hSyn-htTau recapitulates key histopathological features of human tauopathies and represents an attractive model to study the pathophysiology of tauopathies.

METHODS

Experimental animals

All experiments were approved by the State Veterinary and Food Committee of Slovak Republic, and by the Ethics Committee of the Institute of Neuroimmunology, Slovak Academy of Sciences. All animals were housed under standard laboratory conditions with ad libitum access to water and food. Mice were kept on a 12:12 h light-dark cycle. The animals were anesthetized and sacrificed according to ethical guidelines to minimize pain and suffering of experimental animals. We used both C57BL/6 and FVB/NJ mice. The latter were non-transgenic littermates from the rTg4510 line [29] and were used in order to minimize the need for additional breeding. The age at the time of injection is specified below.

AAV vectors

AAV9-hSynapsin1-hTau(151–391/4R)-P2A-mCherry-WPRE (AAV-hSyn-htTau) with a titer of 9.1×10¹³ genome copies (GC) per milliliter (ml) and control virus AAV9-hSynapsin1-mCherry-WPRE (AAV-hSyn-mCherry) with a titer of 6.4 x10¹³ GC/ml were prepared by Vector Biolabs (Malvern, CA). These constructs were chosen because several studies showed that the combination of AAV9 with the human synapsin promotor drives neuron-specific overexpression of mutated tau [10]. Upon delivery, AAVs were aliquoted into 20μl volumes and stored at –80°C. Aliquots that were thawed for experiments were kept in a refrigerator at 4°C or on ice and used for a maximum of 2 weeks. All quantitative experiments were performed and analyzed blind to the AAV vector used (AAV-hSyn-htTau or AAV-hSyn-mCherry). Injection volumes are specified below for each individual experiment.

Electrophoresis and immunoblotting of AAV plasmids

AAV293 cells at 60% confluency were transfected with 5μg of Syn1-htTau plasmid using jetPRIME transfection reagent (Polyplus, 114-07) according to the manufacturer’s instructions. Cells were harvested after 48 h and lysed in 70μl of TTL buffer (20 mM Tris pH 7.4, 150 mM NaCI, 1 mM EDTA, 1 mM DTT, 0.5% Triton X-100, 50 mM NaF, 1 mM Na3VO4, Roche - an inhibitor of protein tyrosine phosphatases, alkaline phosphatases, and ATPases). PAGE 10% gels were loaded with 10μg of protein from lysate from non-transfected cells, 10μg of protein from lysate from transfected cells, 2.5μl (20 ng/μl) of recombinant full-length (2N4R) tau, 2.5μl (20 ng/μl) of recombinant htTau (151–391/4R), and 10μg of protein from lysate from cells transfected with an unrelated Syn1-Synuclein control plasmid. Membranes were blocked with 5% (w/v) milk TBST blocking solution (20 mM Tris-HCL, pH 7.4, 0.15 M NaCl, and 0.1% Tween 20) for 1 h at room temperature and subsequently incubated overnight at 4°C with pan-tau antibody DC25 (Table 1). Membranes were then incubated for 1 h at room temperature with anti-mouse IgG secondary antibody (DAKO) in 5% (w/v) milk in TBST (dilution 1:3000). The blots were developed with Supersignal West Pico Chemiluminescent Substrate (Pierce) and visualized by Luminescence reader LAS 3000 (FUJI Photo Film, Japan).

Table 1

List of primary antibodies used in this study

Antibody/epitope	Host species	Dilution	Source
HT7 (tau aa159–163)	Mouse	1:500	Thermo Scientific (IL, USA)
AT270 (tau pT181)	Mouse	1:500	Thermo Scientific (IL, USA)
AT8 (tau pS202/pT205/pS208)	Mouse	1:500	Thermo Scientific (IL, USA)
AT180 (tau pT231)	Mouse	1:500	Thermo Scientific (IL, USA)
DC18(tau aa168–181)	Mouse	1:1^*	Axon Neuroscience (Bratislava, Slovakia)
DC25 (tau aa347–353)	Mouse	1:1^*	Axon Neuroscience (Bratislava, Slovakia)
DC39C (tau aa434–441)	Mouse	1:1^*	Axon Neuroscience (Bratislava, Slovakia)
mCherry	Rabbit	1:500	Abcam (ab167453)
Neun (neuronal marker)	Rabbit	1:1000	Abcam (ab104225)
CNP (oligodendrocyte marker)	Mouse	1:500	Merck (MAB1580)
GFAP (astrocyte marker)	Rabbit	1:1000	Abcam (ab7260)
IBA1 (macrophage marker)	Rabbit	1:500	Wako (Tokyo, Japan)

^*Supernatant from hybridoma was used.

Stereotaxic application of AAV

Anesthesia was induced by 3% isoflurane, after which the mice were mounted in a stereotaxic apparatus (Robot Stereotaxic, Neurostar). The isoflurane was then kept at 1% to maintain optimal levels of anesthesia. The local anesthetic lidocaine was injected under the skin of the head to minimize pain at the site of the incision. For each AAV we used a dedicated 10μl Hammilton syringe (SGE) and the needle was flushed with 70% ethanol and PBS between injections. The following coordinates from bregma were used for dentate gyrus (DG): AP: –2.5; L: +–2; DV: –2.4 (in mm, DV from skull, bilateral). The virus was injected at 250 nl/min with a 3 min waiting time before withdrawal of the needle. In the dosing study (experimental group 1) we used 1μl and in the longitudinal study (experimental group 2) we used 1.5μl. To obtain more widespread transduction throughout the hippocampus the coordinates from bregma were later refined for injections in CA1 and DG: AP: –2.5; L:+–2; DV: –2.0 & –1.4 (in mm, DV from brain surface, bilateral). We also increased the injection time for subsequent experiments to reduce backflow of the virus when withdrawing the needle. To perform injection at two depths, we first lowered the needle to –2 mm and injected 1μl (100μl/min + 5 min waiting period). The needle was subsequently raised 600μm and another 1μl was injected using the same protocol. For entorhinal cortex, we injected 1μl (200 nl/min + 3 min waiting time) in the left hemisphere using the following coordinates (from bregma): AP: –4.7 mm, L: +/–3.6 mm, DV –3.0 mm (from brain surface). The other hemisphere was injected with PBS using the same protocol. After the surgery, the wound was closed with tissue adhesive (Vetbond) and animals were allowed to recover on a heating pad. Rimadyl (5 mg/kg, intraperitoneal) was applied for post-surgical analgesia.

Experimental groups

Dosing study

5-month-old C57Bl/6 mice were injected with AAV-hSyn-htTau in CA1 and the DG with 1μl containing either 10¹¹ GC/ml (n = 16), 10¹² GC/ml (n = 8) or 10¹³ GC/ml (n = 8). Brains were collected and embedded in paraffin for histological analysis of tau pathology at 1, 2, 3, and 4 months post-injection.

Longitudinal study in the hippocampus

11-month-old C57Bl/6 mice were injected with AAV-hSyn-htTau (n = 8) or negative control AAV9-CMV-GFP (n = 2; Signagen SL100840) in DG. Brains were embedded in cryoblocks for free floating histological analysis of tau pathology at 3, 5, 7, and 9 months post-injection. Control brains were collected only at 9 months post-injection.

Injection protocol optimization

4- to 8-month-old FVB/NJ mice were injected with 2×2μl AAV-hSyn-htTau (n = 6) in CA1 and the DG. Brains were embedded in cryoblocks for free floating histological analysis of tau pathology at 3, 4, 4.5, and 5 months post-injection.

Tau propagation

5-month-old C57Bl/6 mice were injected in the entorhinal cortex with 1μl AAV-hSyn-htTau (left hemisphere) and PBS (right hemisphere) (n = 7) or control AAV-hSyn-mCherry (left hemisphere) and PBS (right hemisphere) (n = 3). Brains were embedded in cryoblocks for free floating histological analysis of tau spreading at one week and three months post-injection.

Expression of human truncated tau

9- to 11-month-old FVB/NJ mice were unilaterally injected in DG and CA1 with AAV-hSyn-htTau (n = 3 males, n = 3 females) and negative control AAV-hSyn-mCherry (n = 1 male, n = 1 female). Brains were collected at 1 month post-injection for expression analysis with ELISA.

Histology & biochemical analysis

6-month-old male FVB/NJ mice were injected with AAV-hSyn-htTau (n = 14) and negative control AAV-hSyn-mCherry (n = 9) in DG and CA1. Afterwards, the brains were collected at 5.5 months post-injection and tested for tau pathology using free floating histology and for the presence of sarkosyl insoluble tau. One AAV-hSyn-htTau injected animal was sacrificed 2.5 weeks post-injection to determine neuronal specificity of the virus.

Behavioral analysis & histology

5.5-month-old C57Bl6 mice were injected with AAV-hSyn-htTau (n = 9) and negative control AAV-hSyn-mCherry (n = 9) in DG and CA1. Behavioral testing was performed 4 months after the injection. The brains were collected afterwards for quantification of tau pathology and gliosis.

Tissue collection

For sample collection, mice were overdosed with intraperitoneal administration of zolazepam/tiletamine (Zoletil 100, 100 mg/kg, Virbac) and xylazine (Xylariem, 20 mg/ml, Ecupharm N.V.) in a ratio of 1:1.7. For experiments that only used tissue for histological analysis, we perfused mice transcardially with PBS containing 2% heparin to remove the blood, followed by 3 min perfusion with cold PBS containing 4% paraformaldehyde for fixation. Brains were carefully removed for histological analysis. For experiments where brain tissue was dissected out for biochemical analysis, animals were only perfused with PBS with 2% heparin for 3 min. The hippocampus was dissected from one hemisphere, frozen in liquid nitrogen, and stored at –80°C for biochemical analysis. The other hemisphere was used for histological analysis. All histological samples were post-fixed overnight in PBS containing 4% paraformaldehyde at 4°C. The next day the brains were washed with PBS and cryoprotected by immersion in PBS containing 30% sucrose at 4°C until the brains sank to the bottom of the tube. The brains were then mounted in blocks of cryomedium (Leica) and stored at –80°C for later sectioning with a cryostat (Leica CM 1850).

ELISA to measure expression of human tau

Tissue was homogenized in a cold room with samples on ice. Samples were incubated in 100μl of lysis buffer (20 mM Tris pH 7.4; 150 mM NaCl; 1 mM EDTA; 2 mM DTT; 0.5% Triton X-100; 50 mM NaF; 1 mM activated Na3VO4; 1x protease inhibitors Complete, EDTA-free [Roche Diagnostics GmbH]) and manually homogenized with a pestle. Samples were then immersed for 15 min in liquid nitrogen, thawed by hand, and then kept on ice for 20 min. Samples were then centrifuged for 20 min, after which the supernatant was placed in a new Eppendorf tube. This supernatant was 100× diluted for the ELISA.

Plates were coated for 2 h at 37°C with DC18 (human tau specific) and DC39C (tau C-terminal) antibodies, 50μl/well. Plates were then washed with PBS-T and blocked for 3 h in PBS-T without a lid at room temperature. Human tau (151–391/4R) and mouse tau recombinant proteins were diluted in PBS-T (5μg/ML). A dilution plate was filled with 300μl/well of recombinant protein or sample and serial 3-fold dilutions were made using a multi-channel pipette. 50μl/well from dilution plate was added for incubation to the ELISA plate. After washing in PBS-T, plates were incubated with anti-mouse secondary antibody (DAKO, 1000× diluted in PBS-T) for 1 h at 37°C. After washing with PBS-T, plates were incubated with HRP substrate (TMB One; Kem-En-Tec Diagnostistics, DK) for 20 min. Reaction was stopped with 0.25 M h2SO4 and the absorbance was read at 450 nm using a PowerWave HT plate reader (Bio-Tek).

Extraction of sarkosyl insoluble tau

Extraction of sarkosyl insoluble tau was performed as previously described [30]. Hippocampi pooled from 4 animals were homogenized in buffer containing 20 mM Tris, 0.8 M NaCl, 1 mM EGTA, 1 mM EDTA, and 10% sucrose, supplemented with protease (Complete, EDTA free, Roche Diagnostics, USA) and phosphatase inhibitors (1 mM Sodium orthovanadate, 20 mM Sodium fluoride). Homogenates were centrifuged at 20,000 g for 20 min, the supernatant (S1) was collected, and N-lauroylsarcosine (sarkosyl) was added to a final concentration of 1% and mixed by stirring for 90 min at room temperature. The sample was then centrifuged at 100,000 g for 90 min at 25°C using Beckmann TLA-100.4 (Beckmann Instrument Inc., CA, USA). Resulting pellets (P2) were washed with PBS by additional spin at 100,000 g for 15 min and finally re-suspended in SDS loading buffer to 1/50 volume of S1 and 20μg w/v corresponding to S1 fraction was used for SDS PAGE.

Western blot analysis

Western blotting was performed according to published protocols [30]. Briefly, proteins were resolved using 12% SDS-PAGE gels and transferred onto nitrocellulose membrane. After blocking in 5% non-fat free milk 1×TBS-Tween, blots were incubated with DC25 (Axon Neuroscience, Bratislava, Slovakia) and GAPDH (Abcam, Bratislava, Slovakia, ab8245) antibodies in 5% fat free milk overnight at 4°C. Following washing, blots were incubated with respective secondary antibodies (Dako, Glostrup, Denmark) and developed using SuperSignal West Pico chemiluminescent Substrate (Thermo Scientific, IL, USA) on Image Reader LAS-3000 (FUJI Photo Film Co, Ltd, Tokyo, Japan).

Immunohistochemistry on free floating cryosections

All steps were performed at room temperature unless specified otherwise. Cryosections (40μm) were first post-fixed in 4% PFA for 2 h. Sections were then washed in PBS and individually dipped in ice-cold 80% formic acid for 40 s for antigen retrieval, and then immersed in 1% H₂O₂ to block endogenous peroxidase activity. The sections were then washed in PBS, immersed in PBS containing 0.3% Triton for 10 min for permeabilization, blocked for 1 h in APTUM section block (Diagnostic Technology), and incubated overnight in primary antibodies dissolved in PBS at 4°C (see Table 1 for primary antibody dilutions). Detection of primary antibodies was performed using the Vectastain ABC Kit (Vector Laboratories, CA, USA) according to the manufacturer’s instructions. Sections were thoroughly washed and incubated in the species-appropriate secondary antibodies. Sections were again thoroughly washed and incubated in ABC solution for 1 h. After washing, the signal was developed using the Vector VIP kit (Vector Laboratories). After washing in PBS, the sections were mounted on glass microscope slides and left to dry for at least 2 h. Sections were then rehydrated for 10 min in PBS and subsequently immersed in 70% ethanol (5 min), 96.5% ethanol (5 min), isopropylalcohol (5 min), and xylene (2×5 min). Slides were then coverslipped using Entallan mounting medium (Merck). Samples were imaged using an Olympus BX51 microscope equipped with Olympus DP27 digital camera (Olympus microscope solutions) using 4×, 10×, and 20× objectives.

Immunofluorescence on free floating cryosections

All steps were performed at room temperature unless specified otherwise. Hippocampal cryosections (40μm) were washed in PBS and permeabilized for 10 min in PBS containing 0.3% triton. Sections were then blocked for 1 h in APTUM section block (Diagnostic Technology) and incubated overnight in primary antibodies dissolved in PBS at 4°C (Table 1). The next morning sections were thoroughly washed and incubated in secondary antibodies dissolved in PBS. Slides were mounted with Vectashield mounting medium containing DAPI (Vector Laboratories). For cryosections containing entorhinal cortex, we started by detecting mCherry using a specific antibody. However, due to the strong signal from mCherry alone we observed no difference with or without antibody staining, and, therefore, later omitted this step. Slides were shielded from light as much as possible during the staining process. The rest of the protocol was identical to the one described above. Samples were imaged with Zeiss 710 Laser Scanning confocal microscope (Zeiss) using 10× and 20× objectives. The laser power was kept constant and the gain of the photo-multiplier tubes (PMTs) was optimized for each imaging session. When images were quantitatively compared, the PMT gain was kept the same for all slides.

To estimate the numbers of cells expressing individual markers, labeled cells were counted manually in randomly selected square areas (200×200μm) in the CA1 (Iba1, GFAP) or CA3 (AT8) regions of the hippocampus. Cells were counted when the soma was visible or when a cluster of processes clearly belonged to one cell. To analyze the morphology of microglia and astrocytes, clearly visible and randomly selected cells (three per animal) were selected in the in CA1 region of the hippocampus. Cells were cropped and deblurred with the Fiji plugin Iterative Deconvolve 3D (one iteration). Cell processes were tracked semi-automatically using the Fiji plugin Simple Neurite Tracer. Values detected in each animal (n = 3) were averaged.

Methoxy-X04 staining on free floating sections

Sections were first permeabilized for 10 min in 0.3% Triton PBS and then thoroughly washed in PBS to remove all detergent. Sections were then incubated for approximately 30 s in Trueblack (Biotium) (diluted 1:20 in 70% ethanol in PBS) to reduce tissue autofluorescence. After washing, sections were incubated for 1 h in 10μg/ml Methoxy-X04 (Tocris Bioscience). As a positive control we used slices from the rTg4510 line at 7 months of age, when these mice have widespread Gallyas positive NFTs in the brain [29]. After thoroughly washing the slides, these were mounted and coverslipped with Vectashield mounting medium without DAPI (Vector Laboratories). Samples were imaged with a Zeiss 710 confocal microscope using 10×, 20×, 40×, and 63× objectives.

Open field exploration and novel object recognition

Only male mice were used for behavioral experiments to reduce variation in mouse behavior caused by the menstrual cycle. Mice were habituated to the behavior room for three days before the start of the experiment. All mice were handled daily to reduce handling stress during the experiments. On experiment day 1, mice were placed in a custom-made behavioral enclosure (45×45×45 cm box) opened on top. The mice were left to freely explore the enclosure for 5 min. A camera for collecting behavioral video data (PointGrey Flea3 1.3 MP Mono USB3 Vision, FL3-U3-13S2M-CS) mounted on top of the behavioral enclosure was used to record the position of the mouse. Bonsai software was used to track the position of the mouse in the box [31]. After each experimental session, the mouse was placed back to its home cage. The box was cleaned thoroughly with 70% ethanol to minimize olfactory cues, and ethanol was left to evaporate before placing the next animal in the cage. On experiment day 3, two identical clean objects were placed in preselected locations in the behavioral enclosure. The mice were allowed to explore freely for 10 min (familiarization phase). Both the objects and the box were cleaned as described before. After 4 h, the same mouse was tested by placing it in the same box (test phase). The box contained one object identical to the previous session and one novel object. The locations of the objects in both the familiarization and test phases were randomized (left/right) for each mouse and group.

Total distance traveled, speed, and open field coverage were evaluated for each mouse and each open field session. Novel object recognition was assessed using novelty preference computed as (time spent exploring novel object)/(total time exploring objects)*100.

Y-maze

Custom-made Y-shaped maze consisted of three equal length arms spaced 120° apart, opened on top. Each arm’s dimensions are: 36×8×20 cm (length×width×height). The mice were allowed to explore the Y-maze freely for 10 min and their movements were tracked and recorded with a camera (PointGrey Flea3 1.3 MP Mono USB3 Vision, FL3-U3-13S2M-CS) mounted on top of the maze. Bonsai software was used to track the position of the mouse in the y-maze [31]

Spontaneous alternations, i.e., the mouse entering a different arm of the maze in each of three consecutive arm entries, were evaluated by computing spontaneous alternation % =(number of spontaneous alternations)/(total number of arm entries-2)*100.

Statistical analysis

Unpaired two-tailed Student’s t-test was used to test for differences in means of respective parameters between groups of animals infected with AAV-hSyn-htTau or control AAV-hSyn-mCherry. Where necessary, p-values were corrected for multiple comparisons using Benjamini-Hochberg procedure to keep the false discovery rate < 0.05. Stars in figure panels indicate statistical significance (p < 0.05).

Data are presented as individual values, group means, and bootstrap estimates of 95% confidence intervals of the mean.

RESULTS

AAV-hSyn-htTau induces expression of human truncated tau in neurons

To induce and visualize selective neuronal expression of human truncated tau, we have designed an AAV (serotype 9) vector capable of inducing truncated tau-specific histopathological changes in wild-type mice. AAV9-hSynapsin1-hTau(151-391/ 4R)-P2A-mCherry (AAV-hSyn-htTau) co-expresses non-mutated human truncated tau(151-391/4R) and mCherry red fluorescent protein in a 1:1 ratio utilizing the self-cleaving P2A peptide (Fig. 1A).

Fig. 1

AAV-hSyn-htTau induces neuron-specific expression of human truncated tau and mCherry. A) Adeno-associated viruses (serotype 9) were prepared using plasmids containing truncated human tau sequence (151–391/4R) linked using a P2A self-cleaving peptide to mCherry red fluorescent protein, under the control of human synapsin promotor 1 (hSyn1). AAV9-hSyn1-hTau(151–391/4R)-P2A-mCherry-WPRE-SV40 (AAV-hSyn-htTau) vectors were stereotaxically injected into hippocampal formation of adult wild-type mice. AAV9-hSyn1-mCherry-WPRE-SV40 (AAV-hSyn-mCherry) was used as negative control. B) AAV293 cell lysate did not express human truncated tau in the absence of transfection (lane 1). Transfection with the AAV-hSyn-htTau plasmid (lane 2) resulted in expression of tau with a lower molecular weight compared to the longest isoform of human tau (lane 3) and the same molecular weight as tau(151–391/4R) synthetic peptide (lane 4). No tau was detected in negative control cells transfected with a plasmid for full-length alpha-synuclein. Tau was detected with pan-tau (aa347–353) antibody DC25. C) Neurons transduced with AAV-hSyn-htTau (red, mCherry) expressed human truncated tau protein (green, detected using human tau(aa168–181)-specific antibody DC18). Scale bar shows 50μm. D) AAV-hSyn-htTau transduction induced expression of human truncated tau at levels comparable to the expression levels of endogenous mouse tau protein. Human tau was detected with human tau-specific antibody DC18, and mouse tau was detected with pan-tau (aa434–441) antibody DC39C that does not recognize truncated tau aa151–391/4R. Gray circles show expression ratios in individual mice (AAV-hSyn-htTau n = 6, AAV-hSyn-mCherry n = 2), black circles and error bars show means and bootstrap 95% confidence intervals (CI) of means (AAV-hSyn-mCherry mean = 0.06, CI = 0–0.08; AAV-hSyn-htTau mean = 0.82, CI = 0.71–0.96). E) Expression of human truncated tau protein was neuron-specific. Expression of mCherry (red) co-localized with neuronal marker Neun, but not with markers for oligodendrocytes (CNP), microglia (IBA1), or astrocytes (GFAP). Scale bars show 100μm.

To confirm expression of truncated tau in vitro, we analyzed the lysate of AAV293 cells that were transfected with the AAV-hSyn1-htTau plasmid and detected human tau(151–391/4R) on western blot at the same molecular weight as tau(151–391/4R) synthetic peptide (Fig. 1B).

For in vivo transduction of human truncated tau and fluorophore mCherry, we performed stereotaxic injections of AAV-hSyn-htTau or AAV-hSyn-mCherry (a control virus) in the hippocampus of wild-type mice. Mice between 4–11 months age (see Methods for details) were injected with 2μl of AAV (10¹³ GC/ml) per hemisphere.

Injection of 2μl of AAV-hSyn-htTau covering DG and CA1 induced expression of human truncated tau in infected neurons (Fig. 1C) at levels comparable to the levels of endogenous mouse tau at one-month post-injection (Fig. 1D). Mean ratio of human tau (detected with DC18 antibody) to endogenous mouse tau (detected with DC39C antibody) was 0.82 (bootstrap 95% confidence interval 0.71–0.96) in mice injected with AAV-hSyn-htTau, and 0.06 (CI = 0–0.12) in control mice.

The AAV9 serotype in combination with the neuron-specific human synapsin 1 promotor is widely used to obtain neuron specific expression [10]. Indeed, at 2.5 weeks post-injection of AAV-hSyn-htTau in hippocampus we observed perfect co-localization with neuronal marker Neun (Fig. 1E). No co-localization was observed with other neuronal cell types such as astrocytes (GFAP), oligodendrocytes (CNP), or cells of myeloid origin such as microglia (IBA1) (Fig. 1E). Throughout the study, no infected mCherry-expressing cells were observed that did not have neuronal morphology.

AAV-hSyn-htTau leads to widespread hyperphosphorylation of tau

Abnormal hyperphosphorylation of tau is one of the hallmarks of developing pathology in tauopathies. To detect hyperphosphorylated tau at multiple timepoints after the initial injection we used AT8 antibody (Fig. 2), which recognizes a triple phosphorylation epitope (pS202/pT205/pS208) with the same affinity as paired helical filaments derived from AD-brains, and other combinations of pS199, pS202, pT205, and pS208 with lower affinity [32]. We detected AT8-positive neurons at three months after the injection of 1μl of AAV-hSyn-htTau in DG in both the 10¹² GC/ml and 10¹³ GC/ml conditions. The higher titer led to higher overall infectivity and number of AT8 positive cells and was, therefore, used for all other experiments. Injection of 2μl of AAV-hSyn-htTau (10¹³ GC/ml) in hippocampus induced robust and widespread diffuse AT8-positive signal in both the dorsal and ventral hippocampus after three months. We detected a similar distribution pattern of hyperphosphorylated tau after four months (Fig. 2A), but instead of diffuse staining we observed more dense clusters of AT8 signal in neurons, indicative of cellular deposits of hyperphosphorylated tau, only in animals infected with AAV-hSyn-htTau (Fig. 2B). We did not observe differences in the progression of tau pathology between C57Bl/6 and FVB/NJ mice. However, any possible influence of genotype could be obscured due to differences in age at the time of AAV injection.

We then checked whether tau hyperphosphorylation also extended to other tau phospho-epitopes relevant to AD and other tauopathies at five months after injection of AAV-hSyn-htTau. Neurons were also positive for phosphorylation at epitopes pT181 and pT231 (Fig. 2C). Importantly, we also detected directly human truncated tau using HT7 antibody in the AT8-positive area (Fig. 2C). The overall numbers of AT8-positive cells declined somewhat at five months when compared to three or four months after injection of AAV-hSyn-htTau. The numbers of AT8-positive cells declined further at seven and nine months after injection, with NFT-like neurons detectable in the tissue (Fig. 2D). The neuronal cell bodies were still present but showed a decline in AT8 immunoreactivity. In sharp contrast to mouse lines with hippocampal neurodegeneration (e.g., rTg4510), no obvious decline in hippocampal volume was observed between four and nine months after the injection of AAV-hSyn-htTau (Fig. 2E). Although we cannot exclude the possibility of very mild neurodegeneration in mice injected with AAV-hSyn-htTau, these results suggest that loss of neurons is unlikely to explain the loss of AT8 signal at later timepoints. It is therefore likely that the expression of truncated tau has attenuated progressively after more than five months.

Fig. 2

Widespread tau pathology develops already at four months after AAV injection. A) Widespread accumulation of hyperphosphorylated tau (AT8+) was detected four months after injection of AAV-hSyn-htTau (left). Higher magnification shows neurons with dense AT8 + staining, indicative of neurofibrillary pathology. AT8 + signal was not detected in animals inject with AAV-hSyn-mCherry (right). Scale bars show 100μm. B) AT8 + neurons were detected in areas with mCherry signal in animals injected with AAV-hSyn-htTau. No AT8-positive neurons were detected in animals injected with AAV-hSyn-mCherry. Neurons were counted in randomly selected square areas (200×200μm). Gray circles show individual values (n = 9 in each group), black circle and error bars show the mean (46.22) and bootstrap 95% confidence interval of the mean (37.22–54.52) of AAV-hSyn-htTau infected group of mice, scale bars show 100μm. C) Phosphorylation of tau at other Alzheimer’s disease-related sites in areas with mCherry signal in animals injected with AAV-hSyn-htTau. The same area as Fig. 2B showed expression of human truncated tau as detected with antibody HT7, which is specific for human tau (aa159–163). Phosphorylation of tau was also detected at threonine 181 (p181) and threonine 231 (p231) in adjacent sections. Scale bars show 100μm. D) Left, AT8 + signal was detected in the pyramidal cell layer of CA1 four months after injection of AAV-hSyn-htTau (arrow), but not AAV-hSyn-mCherry. Right, the signal in these cells was diminished at nine months after injection and cells with dense AT8 staining were detected dorsally to the pyramidal cell layer (arrow), where they were not detected at four months after injection. Scale bars show 100μm. E) Clearly visible cell layers (arrows) and no obvious decline in hippocampal volume were observed between four and nine months after the injection of AAV-hSyn-htTau. Scale bars show 100μm.

AAV-hSyn-htTau induces neurofibrillary tangles, seeding of endogenous mouse tau, and production of sarkosyl insoluble tau

To confirm the presence of NFTs, we stained sections of brains of wild-type mice injected with AAV-hSyn-htTau with Methoxy-X04 (Fig. 3). Methoxy-X04 is a derivative of Congo red dye that binds to beta-pleated sheets that are present in fibrils made of aggregated protein. Specifically, this dye was shown to detect amyloid plaques, Lewy bodies, and NFTs in human patients [33]. At four months post-injection we observed robust Methoxy-X04-positive neurofibrillary tangles in the CA1 region of the ventral hippocampus (Fig. 3A), which also contained abundant AT8 positive signal in the same mouse. A subset of cells was double positive for Methoxy-XO4 and mCherry.

Fig. 3

Injection of AAV-hSyn-htTau leads to accumulation of aggregated tau. A) Widespread positive staining with beta-sheet binding dye Methoxy-X04 was detected in mCherry-positive neurons four months after injections of AAV-hSyn-htTau (arrow), but not after injection of AAV-hSyn-mCherry. Sections were from the same animal as in Figure 2A. Scale bars are 50μm and 100μm (overview). B) Seeding of mouse tau five months after injection of AAV-hSyn-htTau (arrow), but not AAV-hSyn-mCherry. Accumulation of mouse tau was detected with pan-tau (aa434–441) antibody DC39C, which does not recognize truncated tau aa151–391/4R. Scale bars show 50μm. C) Accumulation of sarkosyl insoluble tau in hippocampus 5.5 months after injection with AAV-hSyn-htTau (lanes 1–3), but not in non-injected hippocampus from the negative control (lane 4–5). Lane 6 shows the 6 isoforms of human tau. Lane 7 shows sarkosyl insoluble tau extracted from the brain of a patient with Alzheimer’s disease (positive control). Lane 8 shows the soluble fractions from animals injected with AAV-hSyn-htTau. Tau was detected with pan-tau (aa347–353) antibody DC25. Hippocampi from four mice were pooled before extraction. Each lane 1–5 and 8 therefore shows signal from a total of n = 4 mice.

To examine if AAV-hSyn-htTau can induce seeding of physiological tau, we stained brain sections with DC39C antibody (aa434–441). DC39C antibody recognizes the C-terminal of both human and mouse tau, but not truncated tau(151–391/4R) expressed by AAV-hSyn-htTau and, therefore, exclusively recognizes endogenous mouse tau in our model. At five months after injection we observed DC39C-positive neurons with dark inclusions in areas with AT8-positive neurons (Fig. 3B). The morphology of the immunohistochemical labelling resembled pre-tangles or NFTs, similar to intracellular DC39C+ inclusions found in transgenic mice overexpressing human truncated tau(151–391/3R) [28].

An important characteristic of tau pathology in human tauopathies is the insolubility of tau fibrils in detergents such as sarkosyl. We isolated hippocampi from animals injected with AAV-hSyn-htTau, or wild-type negative controls. The six recombinant isoforms of tau and sarkosyl insoluble tau derived from AD patients were used as positive controls. In agreement with our histological findings of aggregated tau in tissue sections, we observed accumulation of sarkosyl insoluble tau in the hippocampus of wild-type mice injected with AAV-hSyn-htTau. The signal consisted of a double band of truncated tau and was very similar to sarkosyl insoluble tau from the brain stem of transgenic mice overexpressing truncated tau [28] (Fig. 3C).

AAV-hSyn-htTau increases the number of microglia in CA1

Neuroinflammation is a key component of human tauopathies and many aspects are recapitulated in tau transgenic animals [8]. To evaluate basic parameters of possible neuroinflammation we compared numbers of microglia and astrocytes between mice injected with AAV-hSyn-htTau and control AAV-hSyn-mCherry (Fig. 4). At four months post-injection staining with microglia marker Iba1 revealed increased numbers of microglia in CA1 (p = 0.028, unpaired two-tailed Student’s t-test); AAV-htTau mean count = 33.78 (bootstrap 95% confidence interval 27.14–40.89), AAV-mCherry mean count = 22.89 (CI 19.78–29.74) (Fig. 4A). The number of GFAP-positive astrocytes, however, was comparable in the two groups (p = 0.491, Fig. 4B); AAV-htTau mean count = 39.78 (CI 34.89–49.38), AAV-mCherry mean count = 37.11, (CI 34.44–40.34). Thus, the human truncated tau-induced pathology is associated with mild microgliosis, but no detectable increase in the number of astrocytes.

Fig. 4

AAV-induced tau pathology leads to microgliosis, but not astrogliosis. A) Top: Quantification of Iba1-positive microglia in CA1 of the hippocampus revealed higher numbers of microglia in areas transduced with AAV-hSyn-htTau, indicative of mild microgliosis. Scale bars show 500μm and 50μm. Number of microglia was higher in AAV-hSyn-htTau mice than in control mice (p = 0.028), AAV-htTau: n = 9, mean = 33.78, CI = 27.14–40.89; AAV-mCherry: n = 9, mean = 22.89, CI = 19.78–29.74. Bottom: Total length of microglia processes (AAV-htTau: mean = 124.56, CI = 100.36–163.05; AAV-mCherry: mean = 111.92, CI = 103.11–126.81), number of primary processes (6.25, 5.70–7.50; 6.08, 5.33–6.86), and branching density of microglia processes (20.69,19.83–22.08; 19.67, 17.58–22.19) did not differ between animals infected with AAV-hSyn-htTau (n = 8) and control animals (n = 8; corrected p = 0.744, p = 0.783, and p = 0.744, respectively). B) Top: Staining with astrocyte marker GFAP showed no astrogliosis in CA1 transduced with AAV-hSyn-htTau (top). Scale bar shows 500μm. Number of astrocytes did not differ between the two experimental groups (p = 0.491), AAV-htTau: n = 9, mean = 39.78, CI = 34.89–49.38; AAV-mCherry: n = 9, mean = 37.11, CI = 34.44–40.34. Bottom: Total length and branching density of astroglia processes were statistically significantly elevated in animals infected with AAV-hSyn-htTau (n = 9), compared to control animals (n = 9; corrected p = 0.004 and p = 0.011, respectively). Total length: AAV-htTau mean = 363.97, CI = 329.40–398.31; AAV-mCherry mean = 269.31, CI = 233.41–294.19; branching density: AAV-htTau mean = 4.83, CI = 4.47–5.16; AAV-mCherry mean = 4.07, CI = 3.80–4.49. The number of primary processes in astrocytes did not differ between the two experimental groups (corrected p = 0.314; AAV-htTau mean = 6.67, CI = 6.11–7.37; AAV-mCherry mean = 6.26, CI = 5.96–6.82). In A and B, Iba-1 or GFAP-positive cells were counted in randomly selected square areas (200×200μm). Other parameters were evaluated as means of values from three randomly selected well-isolated Iba-1 or GFAP-positive cells in each animal. Gray circles correspond to values from individual animals, black circles and error bars show the mean of each population and bootstrap 95% confidence interval of the mean. Stars indicate statistically significant (p < 0.05) difference between the means.

Despite the increase in the number of microglia in AAV-hSyn-htTau-injected mice, microglia morphological parameters were unchanged compared to control animals injected with AAV-hSyn-mCherry (Fig. 4A bottom). Total length of microglia processes (AAV-htTau mean = 124.56, CI 100.36–163.05; AAV-mCherry mean = 111.92, CI 103.11–126.81), number of primary processes (6.25, 5.70–7.50; 6.08, 5.33–6.86), and branching density of microglia processes (20.69,19.83–22.08; 19.67, 17.58–22.19) did not differ between the two populations (corrected p = 0.744, p = 0.783, and p = 0.744, respectively). On the other hand, although the numbers of GFAP-positive astrocytes were unchanged, the total length of their processes and branching density appeared higher in mice injected with AAV-hSyn-htTau compared to control mice (corrected p = 0.004 and p = 0.011, respectively; unpaired two-tailed Student’s t-test followed by Benjamini-Hochberg procedure, Fig. 4B bottom). We detected an increase in total length of astrocyte processes (AAV-htTau mean = 363.97, CI 329.40–398.31; AAV-mCherry mean = 269.31, CI 233.41–294.19) and their branching density (AAV-htTau 4.83, 4.47–5.16; AAV-mCherry 4.07, CI 3.80–4.49). The number of primary processes in astrocytes did not differ between the two experimental groups (corrected p = 0.314; AAV-htTau 6.67, CI 6.11–7.37; AAV-mCherry 6.26, CI 5.96–6.82).

AAV-hSyn-htTau induces spreading of human tau

Spreading of tau protein between brain cells is a characteristic feature of neurodegenerative process [7]. Co-expression of human truncated tau and mCherry enables visualization of the pool of cells initially infected with AAV-hSyn-htTau. Cells containing human tau without mCherry, however, are those that received tau from other cells, most likely via cell-to-cell transmission (Fig. 5A).

At three months after injection of AAV-hSyn-htTau in the entorhinal cortex, we observed robust expression of mCherry and human tau (HT7 immunoreactivity) (Fig. 5B). In mice injected with AAV-hSyn-htTau as well as in mice injected with the control vector (AAV-hSyn-mCherry), we detected mCherry-positive fibers in the perforant pathway that connects entorhinal cortex with the hippocampus. On the other hand, mCherry-positive cell bodies were rarely observed in the DG, an area directly connected to the entorhinal cortex. At three months post-injection, we detected widespread presence of hyperphosphorylated tau in the entorhinal cortex, consistent with our previous results in the hippocampus (Fig. 5C). Importantly, we have also detected neurons in the entorhinal cortex expressing human tau, but not mCherry, i.e., neurons not infected with AAV-hSyn-htTau, and indicative of local spreading of human tau in the entorhinal cortex (Fig. 5D).

Fig. 5

Propagation of human truncated tau from entorhinal cortex (ECX) to dentate gyrus (DG). A) Schematic representation of spreading of human truncated tau in AAV-hSyn-htTau model. Truncated tau spreads from AAV-hSyn-htTau transduced neuron (for example, in ECX) into neurons in connected areas (for example, DG). B) Overview of a horizontal section (40μm thick) of a mouse hippocampal area at three months after injection of AAV-hSyn-htTau; green: HT7 antibody detecting human tau, red: mCherry, blue: DAPI. Double positive signal from human tau (green) detected with HT7 antibody against human tau(aa159–163) and mCherry (red) was observed in ECX, whereas human tau-only signal was predominant in DG. Cell nuclei were counterstained with DAPI (blue). Scale bar shows 500μm. C) Higher magnification of panel B and comparison with control AAV-hSyn-mCherry. Neuronal fibers positive for mCherry (red) originating from ECX were observed after injection of AAV-hSyn-htTau and AAV-hSyn-mCherry. Strong human tau signal was observed in DG neurons contacted by mCherry-positive fibers only in AAV-hSyn-htTau infected animals, suggestive of propagation of human truncated tau. Scale bar shows 150μm. D) Examples (arrows) of neurons expressing human truncated tau (green) without the expression of mCherry (red), suggestive of inter-neuronal spreading of human truncated tau delivered via AAV-hSyn-htTau, locally in the entorhinal cortex (left) and in the DG (middle, right). Scale bars show 50μm.

Furthermore, at this timepoint human tau was also detected in axons and cell bodies in the DG of mice injected with AAv-hSyn-htTau, but not in those injected with negative control vector (AAV-hSyn-mCherry) (Fig. 5C). Most importantly, virtually all neurons containing human tau were mCherry negative, demonstrating that these cells were not infected with AAV-hSyn-htTau, and suggesting that human tau has spread to these neurons from the entorhinal cortex (Fig. 5D).

Local application of AAV-hSyn-htTau does not induce detectable cognitive impairment

To test if localized AAV-mediated expression of human truncated tau induces major cognitive impairments, mice were bilaterally injected with 2μl of AAV-hSyn-htTau or AAV-hSyn-mCherry. Four months after injections mice underwent behavioral testing in a simple battery of tests, including open field exploration, novel object recognition, and y-maze exploration (Fig. 6).

Fig. 6

AAV-induced tau pathology did not result in detectable behavioral impairments. A) Mice infected with AAV-hSYn-htTau (n = 9) and AAV-hSyn-mCherry (n = 9) covered the same percentage of area in the behavioral arena (left, p = 0.976) and traveled the same total distance during open-field testing (right, p = 0.472). Field coverage: AAV-htTau mean = 3.18, CI = 2.99–3.39; AAV-mCherry mean = 3.19, CI = 2.90–3.48; Total distance: AAV-htTau mean = 31.69, CI = 28.09–36.48; AAV-mCherry mean = 33.72, CI = 30.88–36.57. B) Percentage of spontaneous alternations in Y-maze was comparable in both populations of mice (p = 0.661, n = 9 in both groups). AAV-htTau mean = 55.00, CI = 49.11–62.20; AAV-mCherry mean = 57.44, CI = 51.11–64.81. C) Mice infected with AAV-hSyn-thTau (n = 6) and AAV-hSyn-mCherry (n = 6) displayed the same overall preferences for novel objects (p = 0.598). AAV-htTau mean = 64.56, CI = 44.28–80.11, AAV-mCherry mean = 57.80, CI = 42.76–68.78. Gray circles in all panels correspond to individual animals, black circles and error bars show means and bootstrap estimates of 95% confidence interval of the mean.

We have tested spontaneous locomotion and exploration of mice in an open-field behavioral arena. Both mice infected with AAV-hSyn-htTau and mice infected with AAV-hSyn-mCherry displayed similar extent of the area covered (p = 0.976; AAV-htTau mean = 3.18, CI 2.99–3.39; AAV-mCherry mean = 3.19, CI 2.90–3.48) and the total distance traveled (p = 0.472; AAV-htTau 31.69, 28.09–36.48; AAV-mCherry 33.72, 30.88–36.57) during spontaneous exploration of behavioral arena (Fig. 6A). To assess spatial working memory, we have measured spontaneous alternations in a Y-maze in the two groups of mice (Fig. 6B). All mice have displayed a similar pattern of exploration resulting in the same overall percentage of spontaneous alternations (p = 0.661; AAV-htTau 55.00, 49.11–62.20; AAV-mCherry 57.44, 51.11–64.81). Similarly, we have not detected changes in intermediate-term recognition memory evaluated in a novel object recognition test (Fig. 6C; p = 0.598; AAV-htTau 64.56, 44.28–80.11, AAV-mCherry 57.80, 42.76–68.78).

All evaluated behavioral parameters were similar in both populations of mice four months after the initial AAV injection, indicating that localized sparse expression of human truncated tau did not lead to detectable changes in the cognitive parameters tested in this study.

DISCUSSION

Here we characterized an AAV-based tauopathy model that induces neuronal expression of human truncated tau(151–391/4R). Injection of AAV-hSyn-htTau in the hippocampus of wild-type mice led to widespread accumulation of hyperphosphorylated tau in the entire hippocampal circuit. The accumulation of hyperphosphorylated tau was later accompanied by NFTs and seeding of endogenous mouse tau protein in CA3 neurons. Spreading of human tau from entorhinal cortex to DG was detectable from three months after AAV injection. The developing tau pathology was also confirmed by the presence of sarkosyl insoluble tau in hippocampi isolated from mice injected with AAV-hSyn-htTau, and accompanied by neuroinflammation, as witnessed by a specific increase in microglial infiltration. Although we did not observe an increased number of astrocytes in mice injected with AAV-hSyn-htTau, we did observe specific changes in astrocytic morphology indicative of neuroinflammatory conditions [34]. Hippocampal AAV-induced tau pathology was not associated with detectable changes in several cognitive tests. The development of truncated tau-mediated pathology after application of AAV-hSyn-htTau recapitulates key histopathological features of human tauopathies.

Traditional AAV-based tauopathy models utilize tau protein mutations to increase the propensity of human tau to aggregate [10]. Mutations in the tau gene, however, are not found in AD or sporadic primary tauopathies, i.e., in the vast majority of all cases. In contrast, most tau in NFTs is truncated [35, 36] and most extracellular tau is truncated at both the N- and C-terminals [37, 38] in sporadic tauopathies, including late-onset AD. Pathological tau derived from AD brain efficiently seeds the same tau(151–391/4R) construct used in the current study, suggesting that truncated tau can be efficiently incorporated into tau aggregates [15]. Truncation may additionally play an important role in the pathogenesis of tau aggregation. Tau truncations at the C-terminal, HT368 [39], HT391 [25–28 , 40], or HT421 [41 –43], and the beginning of the proline-rich domain HT151 [25 , 41] lead to a dramatic increase in the toxicity and aggregation potential of tau when expressed in rodents. Transgenic rodent models utilizing the same truncation as the AAV used in this study recapitulate the entire biochemical cascade of tau pathology [25 , 44]. The expression of truncated tau in these models, however, is driven by Thy1 promotor, and strong neurofibrillary pathology is located mainly in the brainstem which is associated with paralysis, i.e., sensorimotor impairment rather than cognitive impairment. In contrast, AAV-hSyn-htTau induces hyperphosphorylated tau and NFTs at the site of injection, for example in the hippocampus.

All AAV-based tauopathy models share some limitations. Most notably, expression of the target protein seems to diminish several months after injection and beyond. Here, five months after application of AAV-hSyn-htTau the mCherry signal was still clearly detectable, staining for human tau, however, showed reduction in signal. NFTs were present, but fewer than would be expected in transgenic animals, for example. This indicates that sustained expression in many neurons is necessary to induce widespread NFTs. We suggest that AAV-hSyn-htTau model of tauopathy is suitable especially for studying early stages of tau pathology.

AAV9 serotype spreads very efficiently throughout the CNS and, therefore, is one of the commonly used AAV serotypes for infecting large numbers of neurons in larger areas [45, 46]. However, the spreading efficiency makes injections into smaller volume areas challenging, for example the targeted injections in entorhinal cortex to study propagation of tau to neighboring cells in the DG. Such difficulties can be resolved by lowering viral titers to infect fewer cells, and by further optimization of injection coordinates dependent of age and genetic background of the animal.

AAV-hSyn-htTau also offers several advantages over traditional transgenic models. Most importantly, AAV-hSyn-htTau can be used to induce tau pathology in every brain region, animal line, and at every age. AAV models overexpressing mutated tau protein were used to study the effects of AD risk genes on tau pathology [47] and tau spreading [48]. In contrast to transgenic animals, AAV models have the advantage that they can initiate tau pathology at any age. Different ages can therefore be compared with the same post-injection interval to study the effect of aging on tau pathology [49] or tau spreading [50]. The induction of localized and targeted tau pathology is particularly useful for studying the mechanisms of tau spreading [51]. The ability to visualize infected neurons using mCherry fluorescence provides stronger basis for the interpretation of propagation of human tau compared to models that require injection of brain extract [11 , 53]. AAV-hSyn-htTau model is also well suited for in vivo studies of progressing tauopathy. Neurons expressing human truncated tau can be identified based on mCherry fluorescence, and their morphology, activity, or function can be compared to nearby healthy neurons within the same area. AAV-hSyn-htTau is therefore a valuable expansion of the toolkit of non-mutated human truncated tau models.

Footnotes

ACKNOWLEDGMENTS

The study was partially supported by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 676144 (Synaptic Dysfunction in Alzheimer Disease, SyDAD), and VEGA #2/0148/16 and 2/0135/18. We would like to thank the following people for their help with experiments: Adriana Murgoci for help with confocal imaging, Veronika Cubinkova and Peter Szalay for help with AAV injections, Vlasta Zahorcova and Samuel Kucko for their help with ELISA, Michaela Skrabanova and Peter Filipcik for their help with plasmid verification, Tomas Smolek, Jana Jergusova, Patricia Karkusova, and Petra Majerova for general troubleshooting and help with histology, Norbert Zilka, Rostislav Skrabana, Branislav Kovacech, and Michal Novak for their input on experimental design.

Authors’ disclosures available online ().

References

Hoglinger

, Respondek

, Kovacs

(2018) New classification of tauopathies. Rev Neurol (Paris) 174, 664–668.

Prince

(2017) Progress on dementia-leaving no one behind. Lancet 390, e51–e53.

Wang

, Mandelkow

(2016) Tau in physiology and pathology.}. Nat Rev Neurosci 17, 5–21.

Novak

, Cehlar

, Skrabana

, Novak

(2018) Tau conformation as a target for disease-modifying therapy: The role of truncation. J Alzheimers Dis 64, S535–S546.

Frost

, Jacks

, Diamond

(2009) Propagation of tau misfolding from the outside to the inside of a cell. J Biol Chem 284, 12845–12852.

Clavaguera

, Bolmont

, Crowther

, Abramowski

, Frank

, Probst

, Fraser

, Stalder

, Beibel

, Staufenbiel

, Jucker

, Goedert

, Tolnay

(2009) Transmission and spreading of tauopathy in transgenic mouse brain. Nat Cell Biol 11, 909–913.

Vogels

, Leuzy

, Cicognola

, Ashton

, Smolek

, Novak

, Blennow

, Zetterberg

, Hromadka

, Zilka

, Schöll

(2020) Propagation of tau pathology: Integrating insights from post mortem and in vivo studies. Biol Psychiatry 87, 808–818.

Vogels

, Murgoci

, Hromadka

(2019) Intersection of pathological tau and microglia at the synapse. Acta Neuropathol Commun 7, 109.

Vargova

, Vogels

, Kostecka

, Hromadka

(2018) Inhibitory interneurons in Alzheimer’s disease. Bratisl Lek Listy 119, 205–209.

10.

Cubinkova

, Valachova

, Brezovakova

, Szabo

, Zimova

, Kostecka

, Jadhav

(2017) Next generation tau models in Alzheimer’s disease research - virus based gene delivery systems. Acta Virol 61, 13–21.

11.

Wegmann

, Bennett

, Amaral

, Hyman

(2017) Studying tau protein propagation and pathology in the mouse brain using adeno-associated viruses. Methods Cell Biol 141, 307–322.

12.

Mondragon-Rodriguez

, Mena

, Binder

, Smith

, Perry

, Garcia-Sierra

(2008) Conformational changes and cleavage of tau in Pick bodies parallel the early processing of tau found in Alzheimer pathology. Neuropathol Appl Neurobiol 34, 62–75.

13.

Derisbourg

, Leghay

, Chiappetta

, Fernandez-Gomez

F-J

, Laurent

, Demeyer

, Carrier

, Buée-Scherrer

, Blum

, Vinh

, Sergeant

, Verdier

, Buée

, Hamdane

(2015) Role of the Tau N-terminal region in microtubule stabilization revealed by new endogenous truncated forms. Sci Rep 5, 9659.

14.

Zhao

, Tseng

I-C

, Heyser

, Rockenstein

, Mante

, Adame

, Zheng

, Huang

, Wang

, Arslan

, Chakrabarty

, Wu

, Bu

, Mobley

, Zhang

Y-W

, St George-Hyslop

, Masliah

, Fraser

, Xu

(2015) Appoptosin-mediated caspase cleavage of tau contributes to progressive supranuclear palsy pathogenesis. Neuron 87, 963–975.

15.

, Jiang

, Hu

, Tung

, Dai

, Chu

, Gong

, Iqbal

, Liu

(2019) Pathological alterations of tau in Alzheimer’s disease and 3xTg-AD mouse brains. Mol Neurobiol 56, 6168–6183.

16.

Courade

, Angers

, Mairet-Coello

, Pacico

, Tyson

, Lightwood

, Munro

, McMillan

, Griffin

, Baker

, Starkie

, Nan

, Westwood

, Mushikiwabo

M-L

, Jung

, Odede

, Sweeney

, Popplewell

, Burgess

, Downey

, Citron

(2018) Epitope determines efficacy of therapeutic anti-Tau antibodies in a functional assay with human Alzheimer Tau. Acta Neuropathol 136, 729–745.

17.

Vandermeeren

, Borgers

, Van Kolen

, Theunis

, Vasconcelos

, Bottelbergs

, Wintmolders

, Daneels

, Willems

, Dockx

, Delbroek

, Marreiro

, Ver Donck

, Sousa

, Nanjunda

, Lacy

, Van De Casteele

, Van Dam

, De Deyn

, Kemp

, Malia

, Mercken

(2018) Anti-tau monoclonal antibodies derived from soluble and filamentous tau show diverse functional properties in vitro and in vivo. J Alzheimers Dis 65, 265–281.

18.

Zhao

, Kotilinek

, Smith

, Hlynialuk

, Zahs

, Ramsden

, Cleary

, Ashe

(2016) Caspase-2 cleavage of tau reversibly impairs memory. Nat Med 22, 1268–1276.

19.

Liu

, Smith

, Huang

, Mahesh

, Vonsattel

JPG

, Petersen

, Gomez-Pastor

, Ashe

(2019) A soluble truncated tau species related to cognitive dysfunction and caspase-2 is elevated in the brain of Huntington’s disease patients. Acta Neuropathol Commun 7, 111.

20.

Smith

, Nelson

, Kemper

, Leinonen-Wright

, Petersen

, Keene

, Ashe

(2019) A soluble tau fragment generated by caspase-2 is associated with dementia in Lewy body disease. Acta Neuropathol Commun 7, 124.

21.

Jeganathan

, von Bergen

, Brutlach

, Steinhoff

H-J

, Mandelkow

(2006) Global hairpin folding of tau in solution. Biochemistry 45, 2283–2293.

22.

Zhang

, Song

, Liu

, Kang

, Kwon

, Duong

, Seyfried

, Hu

, Liu

, Wang

, Cheng

, Sun

, Yu

, Levey

, Ye

(2014) Cleavage of tau by asparagine endopeptidase mediates the neurofibrillary pathology in Alzheimer’s disease. Nat Med 20, 1254–1262.

23.

Liu

, Wang

, Gao

, Dong

, Perry

, Ma

, Wang

(2019) Inhibition of calpain protects against tauopathy in transgenic P301S tau mice. J Alzheimers Dis 69, 1077–1087.

24.

Gamblin

, Chen

, Zambrano

, Abraha

, Lagalwar

, Guillozet

, Lu

, Fu

, Garcia-Sierra

, LaPointe

, Miller

, Berry

, Binder

, Cryns

(2003) Caspase cleavage of tau: Linking amyloid and neurofibrillary tangles in Alzheimer’s disease. Proc Natl Acad Sci U S A 100, 10032–10037.

25.

Zilka

, Filipcik

, Koson

, Fialova

, Skrabana

, Zilkova

, Rolkova

, Kontsekova

, Novak

(2006) Truncated tau from sporadic Alzheimer’s disease suffices to drive neurofibrillary degeneration in vivo. FEBS Lett 580, 3582–3588.

26.

McMillan

, Kraemer

, Robinson

, Leverenz

, Raskind

, Schellenberg

(2011) Truncation of tau at E391 promotes early pathologic changes in transgenic mice. J Neuropathol Exp Neurol 70, 1006–1019.

27.

Filipcik

, Zilka

, Bugos

, Kucerak

, Koson

, Novak

(2012) First transgenic rat model developing progressive cortical neurofibrillary tangles. Neurobiol Aging 33, 1448–1456.

28.

Zimova

, Brezovakova

, Hromadka

, Weisova

, Cubinkova

, Valachova

, Filipcik

, Jadhav

, Smolek

, Zilka

(2016) Human truncated tau induces mature neurofibrillary pathology in a mouse model of human tauopathy. J Alzheimers Dis 54, 831–843.

29.

Santacruz

, Lewis

, Spires

, Paulson

, Kotilinek

, Ingelsson

, Guimaraes

, DeTure

, Ramsden

, McGowan

, Forster

, Yue

, Orne

, Janus

, Mariash

, Kuskowski

, Hyman

, Hutton

, Ashe

(2005) Tau suppression in a neurodegenerative mouse model improves memory function. Science 309, 476–481.

30.

Jadhav

, Katina

, Kovac

, Kazmerova

, Novak

, Zilka

(2015) Truncated tau deregulates synaptic markers in rat model for human tauopathy. Front Cell Neurosci 9, 24.

31.

Lopes

, Bonacchi

, Frazão

, Neto

, Atallah

, Soares

, Moreira

, Matias

, Itskov

, Correia

, Medina

, Calcaterra

, Dreosti

, Paton

, Kampff

(2015) Bonsai: An event-based framework for processing and controlling data streams. Front. Neuroinformatics 9, 7.

32.

Malia

, Teplyakov

, Ernst

, Wu

, Lacy

, Liu

, Vandermeeren

, Mercken

, Luo

, Sweet

, Gilliland

(2016) Epitope mapping and structural basis for the recognition of phosphorylated tau by the anti-tau antibody AT8. Proteins 84, 427–434.

33.

Klunk

, Bacskai

, Mathis

, Kajdasz

, McLellan

, Frosch

, Debnath

, Holt

, Wang

, Hyman

(2002) Imaging Abeta plaques in living transgenic mice with multiphoton microscopy and methoxy-X04, a systemically administered Congo red derivative. J Neuropathol Exp Neurol 61, 797–805.

34.

Perez-Nievas

, Serrano-Pozo

(2018) Deciphering the astrocyte reaction in Alzheimer’s disease. Front Aging Neurosci 10, 114.

35.

Novak

, Kabat

, Wischik

(1993) Molecular characterization of the minimal protease resistant tau unit of the Alzheimer’s disease paired helical filament. EMBO J 12, 365–370.

36.

Basurto-Islas

, Luna-Munoz

, Guillozet-Bongaarts

, Binder

, Mena

, Garcia-Sierra

(2008) Accumulation of aspartic acid421- and glutamic acid391-cleaved tau in neurofibrillary tangles correlates with progression in Alzheimer disease. J Neuropathol Exp Neurol 67, 470–483.

37.

Kanmert

, Cantlon

, Muratore

, Jin

, O’Malley

, Lee

, Young-Pearse

, Selkoe

, Walsh

(2015) C-terminally truncated forms of tau, but not full-length tau or its c-terminal fragments, are released from neurons independently of cell death. J Neurosci 35, 10851–10865.

38.

Sato

, Barthelemy

, Mawuenyega

, Patterson

, Gordon

, Jockel-Balsarotti

, Sullivan

, Crisp

, Kasten

, Kirmess

, Kanaan

, Yarasheski

, Baker-Nigh

, Benzinger

TLS

, Miller

, Karch

, Bateman

(2018) Tau kinetics in neurons and the human central nervous system. Neuron 97, 1284–1298.

39.

Zhang

, Li

, Wang

, Song

, Yu

, Kang

, Liu

, Zhang

, Xie

, Liu

, Wang

, Ye

(2018) Delta-secretase-cleaved Tau stimulates Abeta production via upregulating STAT1-BACE1 signaling in Alzheimer’s disease. Mol Psychiatry, doi: 10.1038/s41380-018-0286-z

40.

Koson

, Zilka

, Kovac

, Kovacech

, Korenova

, Filipcik

, Novak

(2008) Truncated tau expression levels determine life span of a rat model of tauopathy without causing neuronal loss or correlating with terminal neurofibrillary tangle load. Eur J Neurosci 28, 239–246.

41.

Ozcelik

, Sprenger

, Skachokova

, Fraser

, Abramowski

, Clavaguera

, Probst

, Frank

, Muller

, Staufenbiel

, Goedert

, Tolnay

, Winkler

(2016) Co-expression of truncated and full-length tau induces severe neurotoxicity. Mol Psychiatry 21, 1790–1798.

42.

Kim

, Choi

, Lee

, Park

, Kam

, Hong

, Nah

, Jung

, Shin

, Lee

, Choi

, Choo

, Kim

, Choi

, Kayed

, Jung

(2016) Caspase-cleaved tau exhibits rapid memory impairment associated with tau oligomers in a transgenic mouse model. Neurobiol Dis 87, 19–28.

43.

de Calignon

, Fox

, Pitstick

, Carlson

, Bacskai

, Spires-Jones

, Hyman

(2010) Caspase activation precedes and leads to tangles. Nature 464, 1201–1204.

44.

Valachova

, Brezovakova

, Bugos

, Jadhav

, Smolek

, Novak

, Zilka

(2018) A comparative study on pathological features of transgenic rat lines expressing either three or four repeat misfolded tau. J Comp Neurol 526, 1777–1789.

45.

Murlidharan

, Samulski

, Asokan

(2014) Biology of adeno-associated viral vectors in the central nervous system. Front Mol Neurosci 7, 76.

46.

Aschauer

, Kreuz

, Rumpel

(2013) Analysis of transduction efficiency, tropism and axonal transport of AAV serotypes 1, 2, 5, 6, 8 and 9 in the mouse brain. PLoS One 8, e76310.

47.

Zhao

, Liu

, Van Ingelgom

, Linares

, Kurti

, Knight

, Heckman

, Diehl

, Shinohara

, Martens

, Attrebi

, Petrucelli

, Fryer

, Wszolek

, Graff-Radford

, Caselli

, Sanchez-Contreras

, Rademakers

, Murray

, Koga

, Dickson

, Ross

, Bu

(2018) APOE epsilon2 is associated with increased tau pathology in primary tauopathy. Nat Commun 9, 4388.

48.

Audrain

, Haure-Mirande

, Wang

, Kim

, Fanutza

, Chakrabarty

, Fraser

, St George-Hyslop

, Golde

, Blitzer

, Schadt

, Zhang

, Ehrlich

, Gandy

(2018) Integrative approach to sporadic Alzheimer’s disease: Deficiency of TYROBP in a tauopathy mouse model reduces C1q and normalizes clinical phenotype while increasing spread and state of phosphorylation of tau. Mol Psychiatry 24, 1383–1397.

49.

Klein

, Dayton

, Diaczynsky

, Wang

(2010) Pronounced microgliosis and neurodegeneration in aged rats after tau gene transfer. Neurobiol Aging 31, 2091–2102.

50.

Wegmann

, Bennett

, Delorme

, Robbins

, Hu

, McKenzie

, Kirk

, Schiantarelli

, Tunio

, Amaral

, Fan

, Nicholls

, Hudry

, Hyman

(2019) Experimental evidence for the age dependence of tau protein spread in the brain. Sci Adv 5, eaaw6404.

51.

Narasimhan

, Lee

VMY

(2017) The use of mouse models to study cell-to-cell transmission of pathological tau. Methods Cell Biol 141, 287–305.

52.

Wegmann

, Maury

, Kirk

, Saqran

, Roe

, DeVos

, Nicholls

, Fan

, Takeda

, Cagsal-Getkin

, William

, Spires-Jones

, Pitstick

, Carlson

, Pooler

, Hyman

(2015) Removing endogenous tau does not prevent tau propagation yet reduces its neurotoxicity. EMBO J 34, 3028–3041.

53.

Rauch

, Luna

, Guzman

, Audouard

, Challis

, Sibih

, Leshuk

, Hernandez

, Wegmann

, Hyman

, Gradinaru

, Kampmann

, Kosik

(2020) LRP1 is a master regulator of tau uptake and spread. Nature 580, 381–385.