Temporal progression of tau pathology and neuroinflammation in a rhesus monkey model of Alzheimer's disease

Abstract

Introduction: The understanding of the pathological events in Alzheimer's disease (AD) has advanced dramatically, but the successful translation from rodent models into efficient human therapies is still problematic.

Methods: To examine how tau pathology can develop in the primate brain, we injected 12 macaques with a dual tau mutation (P301L/S320F) into the entorhinal cortex (ERC). An investigation was performed using high-resolution microscopy, magnetic resonance imaging (MRI), positron emission tomography (PET), and fluid biomarkers to determine the temporal progression of the pathology 3 and 6 months after the injection.

Results: Using quantitative microscopy targeting markers for neurodegeneration and neuroinflammation, as well as fluid and imaging biomarkers, we detailed the progression of misfolded tau spreading and the consequential inflammatory response induced by glial cells.

Discussion: By combining the analysis of several in vivo biomarkers with extensive brain microscopy analysis, we described the initial steps of misfolded tau spreading and neuroinflammation in a monkey model highly translatable to AD patients.

Highlights: Dual tau mutation delivery in the entorhinal cortex induces progressive tau pathology in rhesus macaques. Exogenous human 4R-tau coaptates monkey 3R-tau during transneuronal spread, in a prion-like manner. Neuroinflammatory response is coordinated by microglia and astrocytes in response to tau pathology, with microglia targeting early tau pathology, while astrocytes engaged later in the progression, coincident with neuronal death. Monthly collection of CSF and plasma revealed a profile of changes in several AD core biomarkers, reflective of neurodegeneration and neuroinflammation as early as 1 month after injection.

Keywords: Alzheimer's disease; biomarkers; glial cells; nonhuman primates; tau.

FIGURE 1

Quantitative 3D analysis of AT8 and ThioS pathology progression in AAV‐2xTau‐injected animals. (A) A summary of the experimental procedures and longitudinal sample collection performed in this study. (B) Fluorescence photomicrographs illustrating the distribution of NEUN (blue), AT8 (red), and ThioS (green) in the hippocampal formation and entorhinal cortex of experimental animals at 3 and 6 months. Tau pathology progression is highlighted by the increased ThioS staining observed in 6‐month animals compared to 3‐month animals. (C) Representative confocal images of the markers examined to quantify disease progression and the four typical neuronal profiles commonly observed in the analyzed regions: healthy, pretangle, mature tangle, and ghost tangle. (D) Three‐dimensional reconstruction of high‐resolution confocal images was performed to identify and quantify tau pathology progression. Expression of each neuronal profile: healthy, pretangle, mature, and ghost tangles, was calculated and corrected for the 3D volume (mm3) occupied by each region analyzed: CA3/hilus, CA1, subiculum (SUB), left ERC, and contralateral ERC (E). A Graphical summary showing the pattern of distribution of AT8 and ThioS across the left and right hemispheres and between the two‐time points investigated is shown in (F). Scale bar: 200 μm (b), 10 μm (c). *p < 0.05 **p < 0.01 ***p < 0.001, two‐way ANOVA, Sidak's post hoc test.

FIGURE 2

Comprehensive analysis of the biochemical progression of tau pathology in the AAV‐2×Tau monkey model. As abnormal splicing in both 3R and 4R tau isoforms is known to be involved in the development of tau pathology in AD, we investigated the expression of both isoforms and their colocalization with AT8. Quantification was performed in the ERC (A) and in the CA1 region of the HIP (B) after the AAV‐2×Tau injection. To further investigate the profile of tau‐induced pathology in treated monkeys, we analyzed a panel of 10 tau‐associated epitopes detected in the ERC‐HF area (C). The TOC1 antibody was used to target tau oligomers, often found in a punctate pattern over cell somas and proximal dendrites. The Alz50 antibody labels early tau phosphorylation, generating diffuse labeling over the soma and abundant labeling of apical and basal dendrites. The antibody specific for phosphorylated tau at S422 produces abundant labeling of somas, dendrites, and axons. TNT2 marker, which identifies an epitope comprised of amino acids 7‐12 present in AD brains, produced highly specific labeling, including a very strong immunosignal in the soma, dendrites, and dendritic spines (highlighted). Antibody TauC3, which labels an early truncated form of tau, produced strong somatic and basal dendritic labeling. The Tau5 antibody that does not bind to rhesus tau, identified human tau protein originated from the viral vector expression, as opposed to endogenous templated tau. The transition from early and intermediate stages of tau alterations into late tangle‐forming species is characterized by predominant somatic staining with a relatively lower contribution of the dendritic compartment, as visualized through antibodies for phospho‐epitopes PHF1 and AT8. The labeling pattern obtained with antibody MN423, a marker for late tau truncation, reveals neurons with substantial morphological alterations and numerous damaged neurites. Finally, ThioS staining was successful in revealing tangles and extracellular remains of tangle‐bearing neurons. Scale bar: 10 μm, *p < 0.05, two‐way ANOVA, Tukey's post hoc test.

FIGURE 3

Progressive hippocampal atrophy positively correlates with AD‐related fluid biomarkers following AAV‐2×Tau injection. Analysis of tau pathology progression in the CSF was performed in all experimental groups. Different phosphorylated tau epitopes such as threonine 181, serine 199, threonine 231, and serine 396, as well as total tau, were longitudinally analyzed following AAV genetic delivery, reflecting tau pathology accumulation in the CSF as a proxy for AD‐related brain pathology (A) [¹⁸F]APN‐1607, was used to visualize misfolded tau progression after AAV delivery across the different experimental groups (B). Changes in standardized uptake value ratio (DSUVR) for [¹⁸F]APN‐1607 in selected brain regions across both hemispheres are shown in (C). ERC = entorhinal cortex, HF = hippocampus, PPHG = posterior parahippocampal gyrus, AMY = amygdala, FUS = fusiform gyrus, MTG = middle temporal gyrus. *p < 0.05 **p < 0.01, ***p < 0.001, ****p < 0.0001, two‐way ANOVA, Tukey's post hoc test.

FIGURE 4

Unbiased stereological counting demonstrates selective neuronal loss across hippocampal regions following ERC injection. (A) Stereological NEUN counting was performed using DAB chromogenic staining across the left (injected side) and the right (contralateral) hemispheres. (B) Total NEUN+ cell density was calculated for the CA3/hilus region, CA1‐CA2, subiculum (SUB), and left entorhinal (L‐ERC). Scale bar: 20 μm. *p < 0.05 **p < 0.01, two‐way ANOVA, Sidak's post hoc test.

FIGURE 5

Progressive hippocampal atrophy positively correlates with AD‐related fluid biomarkers following AAV‐2×Tau injection in rhesus macaques. Longitudinal representation of neurodegeneration‐related biomarkers in the CSF and plasma of control and experimental animals. A robust increase in NF‐L levels in the CSF and plasma was detected early in the disease progression, and levels stayed elevated in the following months until the endpoint collection. (A) Notably, a robust reduction in BDNF levels was also detected early in disease development in both the CSF and plasma. MRI scans were performed in all experimental groups prior to AAV delivery and at the 3‐ and 6‐month time‐points. (B) Representative images for each animal group. These images were collectively used to calculate overall loss of gray matter volume in the hippocampus relative to the baseline. The degree of severity was visually color‐coded by yellow to orange to indicate increasing atrophy, within the white‐squared selected zoom images. (C) Longitudinal volumetric analysis of the hippocampus was calculated for each experimental group, and the correlation of the MRI volumetric changes and NEUN density in the hippocampus, as well as with endpoint NF‐light CSF levels. *p < 0.05 **p < 0.01, ***p < 0.001, ****p < 0.0001, two‐way ANOVA, Tukey's post hoc test.

FIGURE 6

Microglia interacts with early tau pathological markers and drives neuroinflammatory response following 2×Tau delivery. (A) Representative images combining a panel of 10 different markers of tau pathology (red) with microglia marker IBA1 (green), neuronal marker MAP2 (white), and DAPI (blue) across the HF‐ERC region of AAV‐2×Tau injected monkeys at 3 and 6 months. 3D reconstruction analysis indicates microglia is strongly associated with early markers of misfolded tau accumulating in neurons rather than late‐stage markers of the pathology. (B) Fluorescence multilabel microscopy using TNT2 (red), IBA1 (green), and DAPI (blue) was performed and quantified across the CA3/hilus, CA1, and DG regions of all experimental groups. Scale bar: 200 μm and   20 μm. ** p < 0.01, ***p < 0.001, ****p < 0.0001, two‐way ANOVA (B), one‐way ANOVA (D), Tukey's post hoc test.

FIGURE 7

Robust microglial TREM2 expression precedes CA1 hippocampal atrophy. (A) Fluorescence microscopy for the triggering receptor expressed on myeloid cells 2 markers (TREM2, green), the microglial general marker IBA1 (red), and nuclei with DAPI (blue). While TREM2 overexpression precedes tissue neurodegeneration, there is a clear pattern of spatial colocalization between high TREM2 expression and histological damage. (B) Higher magnification of the boxed area in (A) including NEUN (white). (C,D) Representative images of the selected hippocampal areas analyzed and quantification of the 3D cellular density for each cellular marker investigated. A significant reduction in the NEUN+ population occurs concomitantly with a robust increase in total IBA1+ population recruitment. (E) The increase in TREM2 expression by microglia in the brain is associated with an increase in soluble TREM2 levels in the CSF and plasma. Additional analysis of the proinflammatory cytokines TNF‐alpha and IL‐6 was performed in the CSF and plasma across all experimental groups. Scale bar: 500 μm *p < 0.05 **p < 0.01, ***p < 0.001, ****p < 0.0001, two‐way ANOVA, Tukey's post hoc test.

FIGURE 8

Tau pathology progression is associated with microglia and astrocyte‐driven complementary neuroinflammatory responses. (A) Fluorescent photomicrographs illustrating the relationship between glial cells and abnormal tau‐containing cells in the ERC‐HF complex of 3 and 6 M AAV‐2×Tau animals. General markers for neurons (NEUN, blue), microglia (IBA1, green), and astrocytes (GFAP, purple) were combined with the 10 different tau pathology markers previously investigated in this study (red). (A) While microglial interactions predominate for neurons containing early epitopes of tau pathology, astrocytes become the driving force when paired‐helical structures have formed, and neurons progress into the middle‐to‐late stages of the pathology. (B) To further investigate glial cell interaction with tau late‐stage pathology, quadruple staining was performed using PHF1 antibody (Ser396/Ser404) combined with NEUN, IBA1, and GFAP across the ERC‐HF region. (C) Quantification of the total 3D cell density of each cell type across hippocampal regions. The total percentage of neurons expressing PHF1 was calculated based on the total NEUN+ population in the CA3/Hilus, CA1, SUB, and left ERC (C) and reflects a shift in the tangle formation, microglia and astrocyte recruitment across the three and six time‐points period. Scale bar: 50 μm, *p < 0.05 **p < 0.01, two‐way ANOVA, Tukey's post hoc test.