A novel tauopathy model mimicking molecular and spatial aspects of human tau pathology

Abstract

Creating a mouse model that recapitulates human tau pathology is essential for developing strategies to intervene in tau-induced neurodegeneration. However, mimicking the pathological features seen in human pathology often involves a trade-off with artificial effects such as unexpected gene insertion and neurotoxicity from the expression system. To overcome these issues, we developed the rTKhomo mouse model by combining a transgenic CaMKII-tTA system with a P301L mutated 1N4R human tau knock-in at the Rosa26 locus with a C57BL/6J background. This model closely mimics human tau pathology, particularly in the hippocampal CA1 region, showing age-dependent tau accumulation, neuronal loss and neuroinflammation. Notably, whole-brain 3D staining and light-sheet microscopy revealed a spatial gradient of tau deposition from the entorhinal cortex to the hippocampus, similar to the spatial distribution of Braak neurofibrillary tangle staging. Furthermore, [¹⁸F]PM-PBB3 positron emission tomography imaging enabled the quantification and live monitoring of tau deposition. The rTKhomo mouse model shows potential as a promising next-generation preclinical tool for exploring the mechanisms of tauopathy and for developing interventions targeting the spatial progression of tau pathology.

Keywords: model mouse; tau pathology; tauopathy; tissue clearing; whole brain.

Graphical Abstract

Figure 1

Elimination of defects reported in rTg4510 and formation of characteristic tau pathology in rTKhetero/homo mice. (A) Western blotting of human tau and β-actin protein in buffer-extractable fractions from 2-month-old rTKhetero mice (Tau+/−, tTA+/−), Tau-KI responder mice (Tau+/−, tTA−/−) and wild type mice (Tau−/−, tTA−/−). Each genotype n = 3 female. See Supplementary Fig. 10 for uncropped blots. (B) Human tau immunoreactivity labelled by Tau12 antibody in sagittal brain sections from 2-month-old rTKhetero mice, Tau-KI responder mice and wild-type mice. Scale bars = 500 µm. (C) DAPI stained, IBA1 immunofluorescence and GFAP immunofluorescence images of dentate gyrus from wild type and tTA tg on the congenic B6 background, and from wild type and tTA tg on 129 × FVB F1 hybrid background. Scale bars = 100 µm. (D) Areas of dentate granule cell layers in wild type and tTA tg on the congenic B6 background (n = 3 females each), and from wild type and tTA tg on 129 × FVB F1 hybrid background (n = 3 females each). Averaged areas were taken from three serial sagittal sections of each mouse. (E) Ratios of IBA1-positive signals in the dentate gyri. 3 females per group. (F) Ratios of GFAP-positive signals in the dentate gyri. 3 females per group. ****P < 0.0001 (Tukey’s multiple comparison). (G–J) Upper panels show AT8 immunofluorescence staining of hippocampus region from 18-month-old rTKhomo (Tau+/+, tTA+/−) mouse (G) and 6-month-old rTg4510 mouse (H). Lower panels show conformation-specific tau antibody MC1 immunofluorescence staining of hippocampus region from 18-month-old rTKhomo mouse (I) and 6-month-old rTg4510 mouse (J). Inboxes show tau antibody-positive neurons in pyramidal cell layers of CA1. Scale bars = 200 µm.

Figure 2

Whole-brain analysis of tau plaque pathology in rTKhomo mice. (A) A 3D reconstruction and 100 μm projections of both nuclear signal (SYTOX-G) and digitally extracted tau signal (phospho-Tau (AT8)) from coronal and sagittal views of an 18-month-old rTKhomo (Tau+/+, tTA+/−) mouse. The extracted tau signal represents the analysed tau extracted region of interest (ROI), distinguishing it from the background intensity. The data were reconstructed using Imaris software. Scale: 1 mm. (B) Integrated tau signal intensity per regional volume across 53 medium-sized subregions, as defined in a previous report. Integrated tau signal intensity per regional volume was calculated by summing the intensity of tau deposit spots within each anatomical region and normalizing this sum by the total volume of the region (expressed in/mm³). This process was facilitated by the integration of the CCFv3 Allen Brain Atlas annotations and volume metrics through the use of CUBIC-Cloud. The presented value is an average derived from data obtained from three 18-month-old rTKhomo mice. The dotted red line represents the mean density value among the 53 regions. (C) 3D representation and its horizontal, coronal and sagittal views of the spatial tau plot on atlas space from an 18-month-old rTKhomo mouse. This includes 157 subregions within the medium-sized regions that have a tau density above the mean. Each tau plot represents the centroid position of a detected tau ROI label, with the colour indicating the regional tau density rank in descending order, as shown in the colour bar. The upper three sectional views on the right were created using 3 mm projections from each axis in the 3D whole-brain view. The 3D ROI, outlined by dotted boxes and displayed below in 3D and its horizontal, coronal and sagittal views, comprises 47 regions. The lower three sectional views on the right were created using 3 mm projections from each axis in 3D view. Upper scale: 2 mm. Lower scale: 1 mm.

Figure 3

Tau gradient consensus analysis of tau pathology in rTKhomo mice. (A) A scheme of searching tau gradient consensus across grouped regions in the data from the 47 regions within the 3D ROI determined in Fig. 2C from three 18-month-old rTKhomo mice. Regional tau intensity was calculated by taking the sum of total intensity, dividing it by the volume of each region, and then representing the resulting value in log10 scale. The bars are arranged in descending order based on regional tau intensity. This figure illustrates a representative pattern of regions in a conserved order of tau density, with a maximum sequence length of 12 regions, highlighted in magenta. Consensus refers to the common pattern or sequence observed in the tau density gradient across different mice. (B) A representative 3000 μm projection of regional tau density alongside a typical 12-region-sequence tau gradient consensus as highlighted in Fig. 3A. The areas of the regions have been recoloured according to their rank among the 47 regions in descending order. Scale: 1 mm. (C) A network graph illustrating the 12-region-sequence tau gradient consensus. Edge weights were initially determined by adding 1 to the adjacency values among the 256 patterns, and subsequently normalized so that the total sum of all weights equals 250. Colours of the nodes correspond to the rank of the mean regional tau density in descending order. Nodes encircled by a magenta line represent the typical 12-region-sequence tau gradient consensus, as highlighted in Fig. 3A.

Figure 4

In vivo imaging for tau tangle formation using [¹⁸F]PM-PBB3 PET and neurodegeneration in hippocampus. (A) [¹⁸F]PM-PBB3 the standardized uptake value ratio (SUVR) images (cerebellar signal as reference) in 18-month-old rTKhomo (Tau+/+, tTA+/−) and control (Tau+/+, tTA−/−) mice. These images were generated by averaging dynamic data 14–60 min after tracer injection and superimposed on individual MRI T2-weighted images. L, left; R, right; V, ventral; D, dorsal; CR, cranial; CD, caudal. Arrow and arrow heads were signals from non-specific accumulation of [¹⁸F]PM-PBB3 to submandibular gland (white arrow) and Harderian gland (white arrowheads). (B) Time courses of SUVR in hippocampus of rTKhomo and control mice. Data are presented as mean ± SEM. (C) Comparisons of regional SUVR values 45–60 min after tracer injection between two genotypes (rTKhomo, n = 3; control, n = 3). Values are mean ± SEM. Unpaired t test was performed (*P < 0.05). (D) Forebrain right hemispheres were dissected from 3−, 7−, 12−, 15− and 18-month-old mice for western blotting and weighed. There was no significant difference between the two groups. (E) Volumetric analysis of hippocampus of 18-month-old rTKhomo and control mice by MRI T2-weighted imaging (rTKhomo, n = 6; control, n = 6). Values are mean ± SEM. Unpaired t test was performed (*P < 0.05). (F) NeuN immunoreactivities in hippocampus from 18-month-old rTKhomo and 15–18-month-old control mice. Arrowheads indicate thinner CA1 pyramidal cell layer observed in rTKhomo mice. Dotted lines indicate pyramidal cell layers of CA1, CA2, CA3 and dentate granular cell layer (DG). Scale bars = 200 µm. (G–I) Thicknesses of CA1 (G) and CA3 (H) pyramidal cell layers and DG (I) in 18-month-old rTKhomo (male, n = 1; female, n = 6) and 15–18-month-old control (male, n = 4; female, n = 1) mice. Thickness values are mean ± SEM. Unpaired t test was performed (**P < 0.01, *P < 0.05).

Figure 5

Age-dependent glial phenotypic changes in hippocampi of rTKhomo and control mice. (A) Representative AT8 immunofluorescence labelling images in hippocampi of 3−, 7−, 12−, 15−, 18-month-old rTKhomo (tau+/+, tTA+/−) mice. (B) Representative IBA1 immunofluorescence labelling images in hippocampus areas of 3−, 7−, 12−, 15−, 18-month-old rTKhomo and control (tau+/+, tTA−/−) mice. (C) Representative GFAP immunofluorescence labelling images in hippocampus areas of 3−, 7−, 12−, 15−, 18-month-old rTKhomo and control mice. (A–C) Scale bars = 100 µm. (D) Semi-quantification of AT8-positive cell numbers in hippocampus areas of rTKhomo mice. Tukey’s multiple comparisons test was performed (*P < 0.05). Values are mean ± SEM. (E) Semi-quantification of IBA1 signals in hippocampus of rTKhomo and control mice. Values are mean ± SEM with respect to levels measured from samples of age-matched control mice; those levels were scaled to a value of one. (F) Semi-quantification of GFAP signals in hippocampus of rTKhomo and control mice. Values are mean ± SEM with respect to levels measured from samples of age-matched control mice; those levels were scaled to a value of one. 3− (male, n = 3), 7− (male, n = 1 (AT8 = 2); female, n = 2), 12− (male, n = 1; female, n = 2), 15− (male, n = 3; female, n = 3), 18-month-old (female, n = 5) rTKhomo and 3− (male, n = 2), 7− (male, n = 3), 12− (female, n = 3), 15-month-old (male, n = 3), 18-month-old (male, n = 1; female, n = 2) control mice were examined. Since the values for rTKhomo mice are shown as a ratio to the age-matched control group, unpaired t test was performed between groups of the same age (*P < 0.05). (G and H) Scatterplots of IBA1 (G) and GFAP (H) immunoreactivities in hippocampi for AT8-positive cell numbers of rTKhomo mice at 12–18 months of age. Pearson correlation coefficient showed significant correlations (G, P < 0.0001; R² = 0.7329, H, P = 0.0025; R² = 0.5466).

Figure 6

Morphological diversity and spatially tau-dependent response of microglia in the hippocampi of rTKhomo mice. (A and E) Brain sections of 18-month-old rTKhomo (tau+/+, tTA+/−) mice and rTg4510 were immune-stained by IBA1 antibody. Scale bar = 300 µm. (B) Representative images of ramified microglia in dentate gyrus (DG). Arrows show microglial cell bodies. (C and G) Rod-shaped microglia in the stratum radiatum of CA1 region. Arrowheads indicate process tracking of rod-shaped microglia. (D, F and H) Amoeboid-like microglia in CA1 pyramidal cell layer of rTKhomo mice, CA3 and DG of rTg4510. Arrow shows microglial cell body. (I) Quantification of IBA1-positive microglial cell body size in CA1 and DG. 18-month-old rTKhomo (plots include 5339 cells in CA1 and 884 cells in DG from five females; minimum 1003 cells/animal in CA1 and 133 cells/animal in DG, respectively), 6–7-month-old rTg4510 (plots include 2661 cells in CA1 and 1049 cells in DG from two males and three females; minimum 414 cells/animal in CA1 and minimum 117 cells/animal in DG, respectively), 18-month-old control (plots include 2162 cells in CA1 and 293 cells in DG from one male and two females; minimum 535 cells/animal in CA1 and minimum 86 cells/animal in DG, respectively) mice were examined. Box and whiskers plotted Min to Max, show all points. Tukey’s multiple comparisons test was performed on the mean values of each group. (*P < 0.05, ***P < 0.001, ****P < 0.0001). (J and K) Representative AT8 and P2RY12 immunofluorescence labelling images in hippocampi of 15− and 18-month-old rTKhomo mice. The number of AT8-positive neurons was low (J), high (K). Scale bars = 200 µm. (L–O) High magnification images of top images, the region of CA1, the regions of the stratum lacunosum-moleculare and DG. (L and N) AT8 immunofluoresence labeling images. (M and O) P2RY12 immunofluoresence labeling images. (P) Procedure for quantification of AT8 and P2RY12 signals and representative image of region of interest (ROI) for hippocampal AT8-positive and P2RY12 positive signals. Using imageJ software, we set ROIs based on the DAPI signal and quantified the positive area of binarized AT8 and P2RY12 signals. DG, dentate gyrus; SLM, stratum lacunosum-moleculare; SUB, subiculum. Scale bars = 200 µm. (Q–T) Scatterplots of P2RY12 signal (% area) in hippocampal regions for AT8 signal (% area) of rTKhomo mice at 15–18 months of age. 15-month-old (male, n = 6; female, n = 3), 18-month-old (female, n = 4). Pearson correlation coefficient showed significant correlations (R, P = 0.0002; R² = 0.7399, S, P = 0.0104; R² = 0.4634).